- 1

ONE of the characteristics of the people of Tuvalu (the Ellice Islands) is a tendency to depart from scenes of trouble and personal shame. Even at the present day, a man who is not bound by the close ties of family and of property will sometimes take a canoe and a few coconuts and trust himself to the mercy of the ocean rather than endure the reproach of his fellows. In the uncertain currents of the atoll groups this is virtual suicide; and although it is tempered by an element of chance, the primary intention of the one who thus sets out is to sever himself from all past connections even by death. Often, one or more sympathetic companions will accompany him, and their intention in thus wilfully sacrificing themselves is to bring repentance upon the persons responsible for their friend's grievance. In many cases, the departure of young folks is deliberately undertaken in order to punish older relatives for some real or fancied misuse of their authority.

Traditions show that this trait was not uncommon throughout Polynesia, and I would stress the very child-like mental attitude which underlies it, because I think it may have been the primary cause of many of the Polynesian settlements on isolated and unattractive islands, and also - 2 because it offers an explanation of the apparently insufficient reasons given in tradition for some of the more famous voyages. 1

Tradition fails to reveal the cause of the original migration to Vaitupu. In common with most of the atolls in the vicinity, Vaitupu has a legend which makes Samoa the original home of the race, but which simply states, with a distressing lack of detail, that Telematua, the founder, came in a canoe, unnamed, and settled on the island, which he found uninhabited. One might confidently assume that it could not have been the natural advantages of Vaitupu that led Telematua to settle there, for in those days there could have been nothing more attractive on such a lagoon-island than the usual coconut-jungle and a fair abundance of fish in the surrounding waters.

Another legend, imperfectly remembered, and discredited by the present generation, makes Vaitupu the first of all created lands, and its life the first life.

It has been shown by Grimble, 2 on the evidence of Gilbertese traditions, that fragments of that Tonga-fiti host which was driven out of Samoa in the thirteenth century, over-ran the islands of the Gilbert Group and it has been suggested (ibid) that the islands of the Ellice Group formed natural stepping-stones for that migration. The genealogical table of Vaitupu, however, establishes the date of original settlement of that island at about four centuries ago, i.e., some three hundred years later than the expulsion of the Tonga-fiti from Samoa. This is perhaps a matter of small importance, but the fact is worthy of being noted that the - 3 most fertile island of the Ellice Group was insufficiently attractive to any part of that swarm of home-seeking refugees. And it is indeed strange when it is considered that the Gilbert atolls, where many of them did finally settle, are, on the whole, much poorer in all natural advantages than those of the Ellice Group. The only explanation that suggests itself is that the Gilbert Islands were already inhabited by a people of proto-Polynesian stock, who, it has been assumed (loc. cit.), were descended from the ancestors of the Tonga-fiti themselves, and that the lands were, therefore, in a state of cultivation.

This brings us back to the suggestion that Vaitupu was originally settled accidentally by haphazard parties, probably from Samoa, drifting before the south-east Trade: and this point will receive further support from the evidence of pre-European culture.

Many words in old songs, the only reliable examples of the ancient dialect, show a linguistic relationship with Tonga, as distinctly opposed to the Samoa relationship. This may be the result of frequent visits of marauding and slave-seeking parties from Tonga, which are several times mentioned in the legends.

The general characteristics of these isolated Vaitupuans are essentially Polynesian, although their material culture and social organization are of a very low order of development. Their houses, canoes, tools, weapons, and textile manufactures indicate the simplicity and utility requisite to an ease-loving maritime people living in a moderately peaceful community, in no imminent danger from external enemies, and fearing nothing, unless it may be the infrequent famine caused by the failure of the seasonal rains. There is little attempt at decoration or ornamentation, and with the exception, perhaps, of their poetry, the fine arts are represented by the crudest beginnings. Wood-carving, ornamental shell-inlay, feather-work, fine-woven mats, decorative painting, tattoo, and the other arts and crafts usually associated with Polynesian culture had either no beginnings, or such immature development that they were readily swamped by the foreign innovations of later times. A canoe was esteemed for its size, sea-worthiness, and speed, and small - 4 trouble was taken with the finish of its superstructure. A house was a shelter and had attached to it none of the sentiments of home, except insofar as it may have been the burial-place of a renowned ancestor, in which case sentiment was affected by animistic belief.

Manufactured articles of personal adornment had not attained to a state where they could compete for popular favour with the green leaves of the forest, poor bush-flowers, and shells of the reef. Of these last, the pearl-shell was indeed pregnant with poetical meaning, and the peculiar metaphors to which its qualities gave rise will be dealt with under the heading of poetry.

We are not concerned with the history of European discovery and subsequent developments except insofar as they affect the present investigation. Of the few competent observers who have chanced to visit the Ellice Group, none, with the notable exception of the several scientific expeditions to Funafuti, has left us any reliable record of the ancient life and customs. Of the traditional history of Vaitupu, a brief outline will suffice for introduction to the succeeding notes.

Telematua, as previously stated, making a journey from Samoa found Vaitupu uninhabited and settled there. His son was Foumatua and his grandson, Silanga. Silanga made a journey to Tonga, and taking to wife a Tongan woman, Kalasipa, returned with her to Vaitupu.

This is the genealogical table from Telematua to five of the more important families of the present day.

- 5

Allowing 25 years to a generation will give a total of 350 years, thus fixing the advent of Telematua at some time in the fifteen-hundreds.

Another family-group traces descent by another line from Silanga, grandson of Telematua, and the Tongan woman Kalasipa, his second wife.

- 6

This makes 18 generations from Telematua and would thus set back the date of settlement by about a century.

There are stories of inter-island journeys, visits from Gilbert Island voyagers, marauding fleets from Tonga in search of slaves, and of the exploits of various chiefs. Some of these will be referred to under “Traditions.” In more recent times, there are garbled accounts of stray foreign settlers and black-birding raids. About 65 years ago, the first missionaries (Samoan pastors of the London Missionary Society) arrived in the group. By these, the whole population was, without difficulty, induced to accept Christianity. The pastors almost at once assumed autocratic power in the regulation of life and customs. That dialect of Samoa into which the Bible has been translated became the official language of the Ellice peoples. The result of this on Vaitupu, of which island the original dialect bore a much stronger resemblance to the dialects of southern and western than to those of nuclear Polynesia, has been the development of a hybrid patois quite incapable of transmitting the ancestral traditions and folk-lore. 3 These first pastors suppressed also all the old games, dances, and poetry, as well as the customary tattoo and ear-lobe-distension practices of ancient times.

We are not concerned to determine whether the enforced alteration of speech and customs has been to the advantage or otherwise of the islanders, but rather to indicate the factor which has been principally deterrent to ethnographic research. It is indeed only by the liberal application of twist tobacco that the old men can be induced to speak, somewhat - 7 shamefacedly, of the ways of the pouliuli (dark ages) as the term is for the unenlightened times before the advent of the Christian Samoan.

The following notes deal then with a Polynesian community, living in a state of comparative isolation, over a period of 300-400 years, a community which, it may fairly be assumed, was founded accidentally, brought few previous racial traditions to bear on its development, and in which was evolving a primitive culture of local origin—a fact strikingly evident in the nomenclature of the lunar divisions of the seasons.

In the interest of accuracy, it should perhaps be stressed that these notes deal with the island of Vaitupu only, and, although many of the statements might apply equally well to the rites and customs of the peoples of other islands within the group, no claim is made that they represent the Ellice peoples as a whole.

- 8 Page is blank

- 9

1st day	Sae ki tangata—New moon. (Sae: appears—something that has been lost.)
2nd day	ma lua
3rd day	ma tolu
4th day	ma fa
5th day	ma lima
6th day	ma ono
7th day	ma fitu
8th day	ma valu
9th day	ma iva
10th day	ma ngafulu
11th day	Fatatea
12th day	Te fou
13th day	Telamanganai (Sunset and Moonrise) (Se lama i te afiafi se lama i te vaveao) (Impossible to go torch-fishing for flying-fish either in the evening or in the early morning.)
14th day	Te uluaki sengia	1st night of disappearance=Sengi.
15th day	Sengia fakalua	2nd night of disappearance.
16th day	Sengia fakatolu	3rd night of disappearance.
17th day	Sengia fakafa	4th night of disappearance.
18th day	Sengia fakalima	5th night of disappearance.
19th day	Sengia fakaono	6th night of disappearance.
20th day	Sengia fakafitu	7th night of disappearance.
21st day	Sengia fakavalu	8th night of disappearance.
22nd day	Sengia fakaiva	9th night of disappearance.
23rd day	Sengia fakangafulu	10th night of disappearance.
24th day	Tuputasi	1st day of moon in the morning sky.
25th day	Tupulua	2nd day of moon in the morning sky.
26th day	Tuputolu	3rd day of moon in the morning sky.
27th day	Tokatulituli	Sun following (tuli) the moon.
28th day	Tuatua	Sun and moon together.

(1). Mulifa	Ko tena uinga e vale manu mo tangata.
(2) Takaonga	Tena uinga e fakaongaonga te matangi e fakaongaonga tangata.
(3) Kelekele	Tena uinga e angi te matangi e afu to kele to i te fakai.
(4) Silinga	E faka mataku e lasi tena lau matangi.

- 10

(5) Utuaemua	Tena uinga a ia e sili i te matangi e sili i tai.
(6) Toeutua	Tena uinga e faka toka mea katoa ne kosu ne masina ko tea atu masina o te afa.

(1) Fakafu	Tena uinga ko tupu mai fuatanga o lakau ika ko fua matua fonu ko olo ki uta.
(2) Kaunuunu	Tena uinga e tupu taua e loto masei tangata kise tangata.
(3) Luamanu	Tena uinga ko tali ofanga a manu.
(4) Ngataitokia	Tena uinga ko tupu telau matangi, ko tokia tamaliki.
(5) Palolomua	Tena uinga ko lei te au i tai ko lei mea katoa ne masei i masina ko tea atu.
(6) Toepalolo	Tena uinga ko asu te palolo i tai. Mate te matangi. Ko lei mea katoa i tai mo uta i te laukele.

6 moons from West Wind; 6 moons from beginning of Trade Wind.

1st month—Mulifa (Nov.-Dec. moon—the star Mataliki appears in East and Melemele sets in West. (te afa i.e. stormy season of the West Wind commences.)
2nd month—Takaonga	(Windy)
3rd month—Kelekele	(Wind and dust, i.e. kele.)
4th month—Silinga (Superlative gerundive) (Here the Island of Nukufetau interposed 5th month—Silinga-ma.)	(Month of fearful winds and seas.)
5th month—Utuaemua	(Worst winds and seas.)
6th month—Toeutua	(Month of cleaning up and tidying after the months of disturbing winds.)

Ist month—Fakafu	Turtles, fish and trees begin to bear. Roe found in the Bonita—eggs found in the Turtle. This month marked also by the first appearance of the star Melemele at sunset (in the east) while the star Mataliki sets.
2nd month—Kaunuunu	Month of jealousy and tendency of men to fight.
3rd month—Luamanu	Birds build nests.
4th month—Ngataitokia	Month of epidemic sickness (like a gust of wind) children die. (Tokia—euphemism for die.)

- 11

5th month—Palolomua	End of troublous months. The palolo fish appears.
6th month—Toepalolo	The palolo is caught. The wind dies—all's well on land and sea.

NOTE—This gives only 12 lunar months and is probably the cause of the fact that there were two sides of the island in old times involved continuously in heated argument, one side being a month ahead of the other in its reckoning.

A reckoning from Nukufetau Island gives Silinga, then Silingama—making 7 months in the Season of Westerly Winds, and therefore the correct total of 13 lunar months for the year.

te ata po ko fanaka	just before dawn.
soponga o te la	sunrise.
tafa taeao	morning.
fangainga o lupe	about 9 a.m. (lit. feeding-time of pigeons—e.g. Koulua onoono ki te la ko nofo i te fanga-iuga o lupe. Ke olo tatou o tapa aka tou alo—Look you at the sun—it is about 9 o'clock. Let us go and have our meal.)
otea	mid-day (exact).
Tu tonu	mid-day and afternoon.
te la ko pale	early afternoon.
ko singa te la	early afternoon.
afiafi	late afternoon.
afiafi-po	evening.
tonga o te la	sunset (Ko to te la, te kau ata ko lelei ka taeao nei ka kasa te la. The sun is setting. The sunset colours are favourable and so it will be a sunny day to-morrow.)
po	night.
ko vaelua te po	midnight.
valu-a-po	late at night (round about midnight).
taeao	to-morrow.
taeao atu	the day after to-morrow.
ma po	to-night (speaking during the day).
te po nei	last night (speaking any time from dawn till evening).
anafi	yesterday.
anafi atu	the day before yesterday.
tela ko to aso atu	a day before that (indefinite).

Time at which a thing was done, or was to be done, was indicated by showing the slant of the sun with the hand,

- 12

I WAS highly gratified to receive through the kindness of Dr. E. W. Gudger, a copy of his most interesting paper on the distribution of the wooden Ruvettus hook of the Pacific. 4 His introductory statements, and the prefatory remarks of Dr. Clark Wissler on the increasingly recognised importance of fishing methods and implements in the ethnography of the people of the South Seas, have led me to revise a somewhat superficial account I had prepared of some of the maritime activities of the people of Vaitupu, whose population is approximately 600. In this connection, with regard to fishing for palu (Ruvettus—Castor oil fish), I have been able to make a fairly complete survey of the implements employed and method in general use.

The account of this survey will open with a description of the making of lou-palu (Ruvettus hooks) by an old fisherman. Next will come a brief examination of eight selected hooks at present in use on Vaitupu, in order to show individual and family variations. The second section will be devoted to a description of a typical palu-fishing expedition, and the final section will contain some general information an the palu and its purgative effect.

I negotiated with an old tautai (fishing-canoe captain) to make, in my presence, a number of lou-palu, and to ensure careful and correct workmanship I arranged that he should tau-palu (go palu-fishing) with me, with the same hooks when finished. We set out for a tidal flat about five acres in extent and covered with mangrove and ngie (Pemphis acidula) which latter attains there a maximum height of about 15 feet. Beginning with the shrubs near the lagoon, we penetrated towards the centre of this natural shrubbery, examining in all about 60 forked branches, and cutting seven - 13 large ones and nine small ones; the large ones, (see figures), for making the kau-lou (body of the hook), and the small ones for the mata-lou (barb). In selecting branches for making lou, the natives classify forks into three kinds:—

• (1) fāpuku, a fork formed solidly. If the bark is scraped away, the crotch of the fork has a U shape in contradistinction to the V shape of the other types.
• (2) tā-tafai, a fork in which the branch seems to lie along the main stem. The junction is weak and will not resist a separating force applied to the legs. I have applied a force of at least 60 lbs. to the extremities of the legs of even slender specimens of the former (fā-puku) type, in a direction to separate them, without in any way damaging the fork.
• (3) po-kili, a fork which appears, at first glance to be a fāpuku, but in which the bark penetrates the junction for some distance, forming a cleft and thus materially weakening it. Fig. 1 was a fork of this type. It gave way under a slight separating force applied when it was being trimmed.

In choosing the fork, the first requirement was a narrow angle between the legs. The usual branching angle of ngie is about 45 deg. This is much too wide, as it would necessitate a very long (and therefore weak) barb. As a rough check on the space between the legs, the selector in doubtful cases used the length of his index finger in the place where the barb would subsequently lie. In a case where this space exceeded the length of the finger, the fork was rejected. The thickness of the legs was not specifically checked, but it was noticed that the minimum diameter was about 1 in., while the maximum (in the case of Fig. 1, which was subsequently broken) was not less than 2 ins.

Several branches were observed to spiral somewhat from the plane of the longitudinal section through the main stem and the crotch. In such cases, the selector, before cutting the branch, took a one-eye sight in the plane of this longitudinal section, and observed the amount of the divergence caused by the spiral of the branch. If, in his estimation, the divergence was too great to be corrected by a lateral scarf, the fork was rejected. (For examples of lateral scarf-joints, see hooks 1 and 6, Figs 1 to 4.)

- 14

The hooks produced by the members of one family group are usually constant in size for the reason that the makers have a traditional scale of measurement. The tautai (fishing-captain) who was my instructor on this occasion, first chopped, with his adze, about 3 ins. below the fork, nicking the main stem on each side and breaking off the forked head. Then, seizing the smaller branch of the fork in his left hand, little finger in the crotch and thumb extended along the leg, he cut off the head of the branch about 2 ins. beyond the tip of his thumb. Transferring his grip to the main branch of the fork, he severed its head, leaving the leg 1 to 2 ins. longer than its fellow. This measurement, hand's-breadth plus extended thumb, is called apalima. Some families use another unit, slightly larger, te anga—the maximum stretch between thumb-tip and tip of the second finger.

After seven leg-forks had been collected, a search was made for a similar number of barb-forks. This was much more easily accomplished, as forks having an average branch diameter of .7 in. and a branching angle of 45 to 90 deg. were comparatively plentiful. Only one of these deserves mention, (A) of Fig. 1. I had expressed a desire to try out a particular type of hook, often referred to, but apparently not used of recent years. It was described as having a long barb projecting right to the side of the shank leg and spaced laterally from it. It was for the purpose of making such a hook that the fork (A) was cut. The branch of this barb-fork formed an angle with the main stem of somewhat less than 90 deg., and came away from the crotch with a sharp quarter-circle spiral, so that throughout their entire length no part of the two legs lay in the same plane. This fork was subsequently substituted for the fork shown with hook No. 4, Fig. 1, and it is again to be seen in Fig. 3, and finally in the completed form as No. 4, Fig. 4. The legs of all the barb-forks were originally cut roughly to the length of the index-finger. Figs. 2 and 3 show the leg-forks with a tentative arrangement of barb-forks.

The time occupied in selecting and cutting 7 leg-forks and 9 barb-forks was approximately one hour. Before returning to the village, a further 10 minutes was taken up by the tautai in stripping the fibrous skin with its attached row of leaflets from one side of a coconut-frond and weaving - 15 from this a tāpola (rough basket) in which to carry his material.

For the trimming of both leg-forks and barb-forks the only tool used was a light adze made by binding a plane-iron (of the type used in a smoothing plane) to the ordinary Polynesian adze-helve.

The first operation was, holding the fork by the extremities of the legs, to trim the base to the shape required. It was evident that the purpose was to cut away as much superfluous wood as possible without weakening the fork. The bark was scraped carefully out of the crotch and any superfluous soft wood was also cut away so as to give the crotch a U-shape. The fork was then inverted and held by the base while each leg was trimmed in turn. There was no attempt to give the legs an elliptical (cross-section) shape; and in the four hooks shown completed in Fig. 4 the cross-section of any part of any of the legs would be roughly circular. It was subsequently noticed that the cross-sections of the legs of nearly all hooks produced by each individual or family group were constant in general shape and commonly differed from the cross-sectional shapes of those produced by other individuals or family groups.

Two knobs with a saddle between, to be seen clearly on the hooks in Fig 3, were then cut at the back of the head of the longer (i.e. the shank) leg. Next, the space between the head of the barb-leg and the shank was considered with reference to the barb selected. If the distance was greater than the maker's ideal, the scarf against which the barb was to be fitted was cut on the inner face of the barb-leg. If the distance was ideal, or less than the ideal, the scarf was cut on the opposite side (i.e. the front, if the direction of the shank-leg from the barb-leg is called “back”).

Looking at all the hooks under present discussion, of the 7 forks, 6 hooks were completed. Of these, one has a lateral scarf, No. 4. One, No. 7, has the scarf on the inner (back) side of the barb-leg. The remainder have the scarf on the opposite (front) side. Now, that one which has the scarf on the back of the barb-leg has an inside width of 3.3 ins., while for the remainder, the inside widths are 3 ins. or less. This maker's ideal distance is then approximately 3 ins.

- 16

The next operation was to trim the barb-fork knuckle and legs. The smaller branch of this fork, in all cases, became the actual barb of the hook, while the main stem was scarfed to fit against the scarf on the barb-leg of the kau-lou (body of the hook).

It was while observing this operation that I came upon a plausible explanation of the twist remarked on by Dr. Gudger in the setting of the barbs of some hooks. In every scarf-joint, whether inside or outside with respect to the barb-leg of the hook, one scarfed surface must be on the outside of its fork-leg while the surface which it meets must be on the inner side of its respective fork. Now in cutting with an adze to form a scarf-joint flat surface on the inner side of the leg of a fork, the crotch of the fork is held up in the free hand, allowing the extremity of the leg to be cut to rest upon a solid surface. If this action is followed carefully, it will readily be seen that it is impossible to manoeuvre the adze into such a position that it will cut a flat surface at right angles to the plane of the longitudinal section through the legs. In other words, the flat surface will face the handle of the adze, which is, of necessity, slightly to one side of the other leg of the fork. The slanting surface thus cut, when applied to the flat surface on the leg to which it is to be scarfed, gives the barb a twist to one side. It will thus be clear that in hooks made by a right-handed person the tip of the barb will have a tendency to deviate to the left of the medial longitudinal plane of the hook (i.e. looked at from behind the shank-leg), and vice versa. It was noticed that when this effect caused a pronounced twist some effort was made to correct it by paring the faulty surface.

Exhaustive questioning on this point elicited such answers as: “Yes, I tried to give it a twist because it is easier for the palu (Ruvettus) to get on,” or, “ … because it is more difficult for the palu to sao (become unhooked);” or, “Every man follows his own manatu (inclination?) in these matters.”

On the whole, I think these explanations are more often children than parents of the twist.

Of 25 Vaitupu hooks, chosen at random, the barbs of 7 had a deflection to the left (looked at from behind the

- i - ii - 17

shank-leg), 5 had a deflection to the right, and 13 pointed directly at the medial line of the shank-leg.

On all iron hooks, native-made on the same principle as the wooden hook (see Fig. 6 for three typical specimens), which I examined, the barbs pointed directly at the shank-leg. This would seem to give weight to the point made above, viz., that the barb-deflection noticed in some wooden hooks may be accidental.

The final operation was cutting the point of the barb so as to give the required clearance, and trimming it off to the conventional shape.

Here, however, occurs a point which throws considerable light, not only on the deflection of the barb in some cases, but also on the particular action of the hook in taking a palu, and which, therefore, illustrates the theory (to be discussed later in this section) on which the construction of the hook depends. The clearance of the barb-point was estimated in all cases with reference to the thickness of the maker's thumb in the region of the thumbnail. It was noticed, however, that in some hooks he allowed a clearance which would barely accommodate the thumb (point of the barb against the middle of the thumbnail), while in others the thumb could pass through the clearance without touching either the shank-leg or the barb.

In several cases the length of the barb was guessed, and the point cut and finished before the scarf-joint was made. Now in two such cases the maker found that he had over-estimated the length of the barb, since, on fitting the scarf-joint, the point of the barb was brought too close to the shank-leg to allow the insertion of the thumb. In these two cases, he deliberately pared the fitting surface of the scarf-joint on one side so as to twist the barb-point to that side of the shank-leg and so increase the clearance to the requisite amount by the addition of some lateral clearance.

Asked for an explanation of the difference in the amount of clearance allowed in different hooks, he replied with a lengthy exposition of the action of the hook in taking a palu. This will be reserved for consideration later. The following is a summary of his remarks with regard to the point under discussion.

- 18

“Fishing at great depths, it is difficult to feel anything in the nature of a ‘bite.’ It is only after long experience that a man can distinguish between the first long pull of a palu, and the pull of the bare line and sinker as the canoe lifts on the swell.

“Now, when the clearance is wide it is easy for a palu to become hooked, and easy also for it to escape if the fisherman does not at once begin to haul in. On the other hand, if the clearance is narrow, the palu takes a much longer time to force its jaw over the barb, and once on, it is difficult for it to get off again. A fisherman will thus save time if he is sufficiently skilled to operate a hook with a large clearance.”

I have seen inexperienced natives using hooks with a narrow clearance, and attaining a fair success even when they succumbed to the soporific effect of the canoe rising and falling with the swell, and woke up, only at intervals, to haul in sometimes a palu, and sometimes what the sharks had left of one.

I examined the 25 hooks referred to above, testing the barb-clearance with this “rule of thumb.” In twelve hooks the point of the thumb could be passed, edgewise, through the clearance, touching both barb-point and shank-leg simultaneously. Of the remaining 13 hooks, 6 had a clearance not more than .3 in. greater than the thickness of the thumb. The other 7 were abnormal, with clearances varying from 1 to 2.4 ins., and of these, 3 belonged to an expert palu-fisher, two having the barb curled sharply down like a fowl's claw, and 6 of the 7 having the barb-leg of more than average length, a factor which naturally assists the barb in its function of preventing the palu from shaking its head off the hook.

Any pronounced irregularities on the surface of both barb and body were now smoothed off with a rasp, which was a hard dry piece of the skin taken from the lower vertebral region of the back of a species of long-tailed ray.

The next operation was the lashing of the scarf-joint. The barb-leg was held horizontally in the left hand, knuckle uppermost, the first finger and thumb holding the barb in position. The lashing-string was three-plait coconut-sennit braid of an average width of ⅛in. The end of the string was held under the thumb and the lashing commenced, down - 19 on the distal and up on the proximal side of the leg. A deliberate strain was given to each half turn and the thumb was moved down to hold each completed turn as the lashing proceeded.

Each alternate turn was made to cross and seize its predecessor just before the upward strain on the proximal side. The turns thus proceeded in pairs, and the seizing made a more or less pronounced raphe down the front of the barb-leg. The lashing continued until only about ¼in. of the scarf-joint remained visible. A loop was then made, the free end pointing upward along the barb-leg, and the lashing was continued over this free end until two or three turns beyond the end of the scarf-joint. The free end was then drawn up tightly, taking in the remainder of the loop, and cut off.

To assist him, the maker had a younger member of his family whose method was the same in all but three points:—

• (1) He commenced his lashing by taking several figure-of-eight turns around the knuckle.
• (2) He used a figure-of-eight lashing which produced a double layer of string around the joint and a distinct and neat raphe down the front.
• (3) He finished off by catching up the last four turns with the free end of his string, drawing them together and knotting off.

The attachment of the snood was accomplished by teasing out the component strands of the end to be attached for 2 or 3 ins. from the end, placing the lower end of the snood-line closely against the front of the shank-head, directly opposite the depression below the lower knob, and drawing the loosened strands, half on each side around the shank below the lower knob and back to the snood. Around this they were then twisted several times. Next, a length of the same 3-plait binding-string (tuli kafa) as was used to lash the barb-joint, was bound several times over these strands and round the snood. It was then used to serve the snood opposite the lower knob and again seized the snood to the shank in the region of the saddle, the free end finally serving the snood for some distance and being finished off by a method similar to that employed for finishing the barb-joint lashing.

- 20

In three of the six hooks the snood is made of 3-plait braid, sennit, (kafa) of average width, .3 in., and average thickness, .2 in. Two hooks have snoods of 3-ply twist sennit (kafakini) .4 in. diameter. The length of the snood of each of these hooks is given in the table which follows. It is roughly determined, in making, by the length of the forearm. The outer ends of those snoods made of 3-plait braid terminate in loops 2.5 to 3.3 ins. in length. These loops, however, although in one piece with the material of the snoods, are of 2-ply twist. The other three snoods terminate in large double knots. In all cases the snoods taper towards the outer (knot or loop) end. The explanation given for this was that it was easier to tie a fast knot with a thin cord than with a thick one.

Made by a Vaitupu tautai (fishing-canoe captain) in March, 1928.

All measurements are in inches. Outside measurements follow the curve, inside are shortest distance.

The numbers correspond with the numbers on Fig. 3 (Hook No. 1 was broken in making).

Number		2	3	4	5	6	7
Shank-leg	Inside	9.1	8.9	8.0	8.2	8.8	8.1
	Outside	13.0	12.1	11.4	11.4	11.9	11.6
Barb-leg	Inside	7.0	7.8	6.7	6.6	6.1	6.4
	Outside	10.7	11.5	10.7	10.2	9.2	10.5
Barb	Inside	2.0	2.3	3.4	2.3	2.2	1.9
	Outside	3.5	3.8	4.5	3.7	3.4	3.7
	Clearance	.6	.5	.6	.7	.4	.7
Weight in ounces		9.7	6.5	6.8	6.2	6.7	5.9
Max. Width	Inside	2.3	2.7	2.5	3.1	2.6	2.7
	Outside	4.2	4.9	4.4	4.9	4.2	5.0
Length of Snood from lower edge of seizing to outer end of knot or loop.		23.5	18.5	22.1	21.2	18.0	19.5

Now, in order to show family differences, I obtained, on loan, all the palu hooks at present in use among eight distinct palu-fishing family groups. And of these, one was chosen as closely representing what I consider to be the ideal normal hook (see No. 2, Fig. 5) while the other 7 were chosen to show family differences in size and construction. The tables which follow give sufficient data for comparison.

- 21

Three typical iron hooks are to be seen in Fig. 6. These hooks are very popular, and occur with the wooden hooks in the proportion of about 3:2. It is quite probable that, if sufficient iron rod were available, no more wooden hooks would be made. No imported hooks are used for taking palu at Vaitupu, but at Funafuti, 74 miles to the south-east, they are much used.

in actual use on Vaitupu (Ellice Islands) in March, 1928, collected from eight distinct family groups.

All measurements are in inches. Outside measurements follow the curve, inside are shortest distance.

The following numbers correspond with the numbers on Figs. 5, 6.

Number		1	2	3	4	5	6	7	8
Shank-leg	Inside	10.1	7.0	10.1	9.2	9.2	13.7	8.2	7.5
	Outside	13.9	10.0	13.6	13.0	12.6	17.8	10.9	10.1
Barb-leg	Inside	7.5	6.1	7.9	7.0	7.5	8.4	6.2	6.3
	Outside	11.0	9.3	12.1	10.0	11.3	11.8	9.2	9.3
Barb	Inside	1.6	1.5	1.5	2.0	1.7	1.8	1.6	2.3
	Outside	1.6	1.5	1.5	2.0	1.7	1.8	1.6	2.3
	Clearance	1.2	.5	1.0	.9	.4	2.4	.6	.4
Weight in ounces		12.5	6.8	13.6	5.7	6.7	15.2	4.5	5.9
Max. Width	Inside	3.0	2.5	3.0	3.2	2.5	4.2	2.5	3.0
	Outside	5.3	4.3	6.0	5.2	5.2	6.1	4.4	5.2
Length of Snood from lower edge of seizing to outer end of knot or loop.		18.4	15.0	22.8	16.1	13.6	14.5	17.1	16.1

SHAPE.

	LEGS	BASE	BASE
Hook No.	Cross Section	Front-rear Section	Cross Section
1.	Elliptical	U (wide)	U
2.	Elliptical	U (wide)	U
3.	Round	U	V
4.	Shank-leg triangular, apex to rear. Barb-leg round.	V (wide)	V (sharp)
5.	Four-sided with ridges front, rear and lateral.	U	U (narrow)
6.	Square, corners rounded, upper barb-leg circular.	V (wide)	U (flat-based)
7.	Elliptical	U	V
8.	Circular	U (rough)	U (narrow angular)

- 22 BARB.

Number of Hook	Joint with respect to barb-leg.	Lashing.	Approximate angle of lateral medial line to the horizontal.
1.	Scarf inside.	3-plait braid, sennit, raphe in front.	Hooked downward. 45 deg. at point.
2.	Scarf inside.	2-strand twist, sennit, figure-of-eight lashing, raphe in front. One turn over a groove on the upper surface of the barb knuckle.	18 deg. below.
3.	Scarf inside.	3-plait braid, sennit, two layers. Figure-of 8 over groove on upper face of barb knuckle two turns one way and one the other, upper layer is a series of half hitches.	7 deg. below.
4.	Scarf inside.	3-plait braid, sennit, single layer, lower end seized by 3 turns and returned to the knuckle where it makes several turns, one being over a groove in the upper face of the barb knuckle. No raphe.	Hooked down ward. Chord of the arc horizontal.
5.	Scarf outside	3-plait braid, sennit, double layer for inch below the barb. Remainder single layer half hitches, raphe in front. One turn over the upper face of the barb.	Horizontal.
6.	Scarf inside.	33 turns of imported cotton line wound flat. Upper layer of 10 turns slightly spaced.	14 deg. above.
7.	Scarf inside.	3-plait braid, sennit, 2 layers, upper layer made of half hitches. Raphe in front.	Horizontal (upper surface curved and gives point appearance of downward deflection.
8.	Scarf outside	3-plait braid, sennit, single layer, lower end seized under one turn and brought back to knuckle where the binding is finished in three half hitches.	Curved. Inner half to point a bout 15 deg. below.

- 23

Hook No. 1—The snood is two lines of 3-plait braid, sennit, approximately ⅓in. wide and ⅛in. thick. The shank-head has two knobs at the back under each of which each line of the snood has one complete turn. The whole is roughly seized by 5 turns of finer 3-plait braid, sennit, one below the lower knob and 4 in the saddle. The upper end of this braid then seizes the snood-lines for about one inch. The outer end of the snood terminates in a double knot (.7in. diameter). The barb of this specimen has a greater (fowl's-claw) hook than any other found on Vaitupu.

Hook No. 2—The snood is of 3-ply twist, sennit (kafa-kini). Its components are separated and wound around the lower knob. This is then seized by 5 or more turns of 3-plait sennit wound alternately round both shank and snood and between them (in semi-figure-of-8 lashing). The lashing string is then used to serve the snood opposite the lower knob and again seizes it to the shank-head, with four turns round the saddle, of the same kind as those below the lower knob. The remainder of the lashing-string is then used to serve the snood for a distance of 8.1 ins. from the uppermost turn of the shank lashing. The outer end of the snood terminates in a knot made by separating the 3 strands, then again dividing the components of each strand into three and plaiting it into a braid which is knotted around the body of the snood, one strand knotted above the other until a knot of .7 in. thick and 1 in. long is produced.

This hook approaches what I consider to be the ideal of a wooden Ruvettus hook. It is strong, light, well-balanced, and neatly finished. The barb points directly at the shank, indicating careful workmanship in the scarf-joint. The barb-leg has a longitudinal crack on the left side, but it does not seem to have impaired the strength of the hook as it is still used. There are many teeth-marks on the barb-leg on both inner and outer surfaces.

On the whole, I think, if a composite picture were to be made of all wooden hooks at present in use on Vaitupu, it would closely approach this specimen in both shape and size. If it is taken as a norm from this point of view, the other 7 hooks, chosen principally for their eccentricities, will readily show the individual and family variations possible on one small island.

- 24

Hook No. 3—The snood is of 4-plait sennit (kinalolo) and terminates in a single knot at the outer end. It is fastened below the lower knob by a double-half-hitch. The whole is then seized by 12 turns of 2-ply twist sennit (kolokolo). This roughly but firmly envelopes both snood and shank-head, two turns being round the double-half-hitch and the remainder in the region of the saddle. The free end serves the snood for a distance of 2.2 ins. above the lashing. The end projecting downwards from the double-half-hitch has a single large knot to prevent its becoming un-hitched.

On the whole, this is the roughest and perhaps also the strongest, of the 8 hooks. Some of the original bark still adheres to both legs.

Hook No. 4—The snood here is secured below the lower knob by a single half-hitch, the lower end of which is twisted once around the main turn and brought up parallel with the snood for a distance of one inch. It is seized below the lower knob, and in the saddle by 5 and 7 turns respectively of imported cotton fishing-line. The outer end of the snood ends in a loop 2.2 ins. long.

Hook No. 5—The snood is of imported cotton-line. The seizing is of fine 3-plait braid (tulikaf a) lashed figure-of-8 style.

This hook has an eccentric barb. The knuckle has a blunt forward projection of .6 in. It has few teeth-marks, but has blood stains and other marks of considerable usage.

Hook No. 6—The fashioning of the cotton snood of this hook is clear in the figure. There is no seizing.

The hook is the largest found on Vaitupu. The barb-clearance is the greatest I have ever seen. The barb-leg has teeth-marks. The shank-leg tapers towards the top and has three undercut knobs, the lowest of which is the largest and the topmost the smallest. The distance between the ridges of lowest and middle knobs is 1 in. and between those of middle and topmost knobs .5 in.

As shown in the table, the base has a flat-bottomed, U-shaped cross-section. Two other hooks by the same maker (not included in the eight here described) have this same eccentric base.

Hook No. 7—This is one of the frailest hooks found, but it has more undoubted teeth-marks on both legs than any other hook examined. Owing to the slimness of the - 25 legs the barb has a springy lateral movement when pressure is applied. This, I should imagine, greatly assists the hooking of a palu. The snood is of 3-plait braid sennit, .3 in. wide (kafa). The lashing string is the same material but finer (tuli kafa). The lashing is normal.

Hook No. 8—The snood and snood-lashing are similar to those of No. 7. This hook is eccentric in that, although the angle between the shank and the barb-leg is wide, the barb scarf-joint is on the outside of the barb-leg, thus making necessary a long (and therefore weak) barb.

After the above had been written, a hook came to hand from still another family and, as it differs in two striking respects from any other hook described in this paper, it is here put on record and figured (Text-Fig. 1).

As may be seen by reference to text-Fig. 1, it is a normal Vaitupu hook except in that it has only one knob for the snood, which is made of double imported cotton-line, and in that the barb is more eccentric in shape than any other I have discussed. This barb has been cut so that a projection (part of the main stem below the fork) .4 in. in height, extends above what would be the knuckle on a normal barb, and ends flush with the head of the barb-leg, forming an upward extension of the scarf-joint. Two turns of the joint-lashing are taken around this extension of the scarf,

- 26

The following list gives the native names for all parts of the hook with their English equivalents:—

From a brief classification of the points of difference and resemblance in the measurements and other data given for these 15 Vaitupu hooks the following deductions may be made:—

(1) The shape of the barb, its angle to the horizontal, the shape of the base, the cross-section of the legs, the attachment of the snood to the shank-head, the degree of neatness in finish, and the general size of the hook remain roughly uniform in hooks made by the members of one family; and commonly, though not necessarily, differ from those found in hooks made by other families. In making this statement, I have taken into consideration, not only the six hooks first described, but also other hooks obtained from some of the families who are each represented later by only one hook.

(2) The length of the barb, its twist (if any), the clearance, the position of the scarf-joint, the method of lashing the scarf-joint, the cordage used for the snood, and the termination of the snood (i.e. knot or loop) differ, not only among families and individual makers, but even among hooks made by the same individual.

(3) The only constant is the material of which the hooks are made which is invariably ngie (Pemphis acidula).

Although irrelevant, it is perhaps worth recording here that several natives of the older generation on Vaitupu have informed me that in their young days they saw hooks such as those from Fakaofo, Bowditch Island, described by Dr. Gudger (loc. cit.) with the barb- and snood-lashings protected by a layer of kaka (natural fibre cloth from the base of the coconut-leaf). It was more common however to - 27 smear the lashing with pulukau (gum). The reason for this was to prevent the string from being frayed or cut by the teeth of the palu. The same people declare that some families were accustomed to reinforce the taka (snood) by tying to it a stick roughly equal in length to the snood and about ¾ in. in thickness. This was to prevent the snood's being cut by the sharks which usually attack the palu when it is being hauled up. Also used by some makers in former times, was the ridge-and-groove modification of the scarf-joint.

The construction of this peculiar fish-hook, the development of which to its present variety of forms probably took many hundreds of years, is explained by the native makers with reference to their theory of its action in taking a palu. The following is a summary of the more reasonable opinions offered in this connection.

When the palu, attempting to take the bait, feels the resistance of the line, he wheels about and tries to carry it off in a direction opposite to that resistance, i.e. downward. (Here, it should perhaps be made clear that neither hook nor sinker is ever allowed to approach the bottom.) The fisherman takes the strain on his line but does not haul in. The palu, struggling to overcome this resistance, gradually forces the angle of his jaw or, perhaps, the lower jaw itself, through the clearance between the barb-point and the shank of the hook. Then, finding the bait in his throat and being still unable, either to swallow it or to carry it away, he attempts to free himself of it by throwing his head upward. This action causes the barb-point to penetrate either the gills or the tissues of the floor of the mouth. The fisherman, feeling the release of the first strain, immediately begins to haul in quickly. The whole of the barb-leg then follows the barb and the jaw-bone slips into the crotch between the legs of the hook. Now, at irregular intervals, the palu resists with such vigour that it is impossible to draw in the line more than an inch at a time. At times he lies for several minutes as a dead weight on the end of the line and allows himself to be hauled in steadily, while frequently he swims upward so quickly, that it is difficult to - 28 keep a strain on the line. It is on these last occasions that he would be able to shake his head off the hook if it were not for the transverse barb with its very narrow clearance.

My whole experience of palu-fishing corroborates the above. The native fishermen at Funafuti who use imported hooks strike quickly and haul in. On one occasion a Funafuti fisherman formed one of the crew of my canoe at Vaitupu. I was using native-made iron hooks (see Fig. 6). I noticed that whenever this Funafuti man, who was occasionally given a line to hold, felt the first pressure, he threw his arm up violently in an attempt to “strike.” Although he tried at least half-a-dozen times he was unable to hook a fish. Again, I have noticed that an inexperienced fisherman will sometimes haul in before the first pull of the palu has ceased. On such occasions, the barb may become wedged against the jaw-bone, and if the fish is a large one, either the barb will be broken at the knuckle or the legs of the main fork will part company. I have seen hooks broken in this manner, and the only palu I have ever seen brought to the surface with the barb caught in the angle of its jaw was not more than 2 ft. 6 ins. in length.

Once the palu has been properly hooked (and it is important to remember that the only true indication of this is the sudden cessation of the first long, steady, downward pull on the line) the object of the fisherman is to keep the jaw-bone down in the crotch between the legs. It is for this reason that he never “plays” the fish but hauls in as quickly as may be and, at times of particularly fierce resistance, even bends his line over the side of the canoe, trusting to the strength of his tackle rather than allow the palu to get back up to the head of the barb-leg where he might, by his struggles free himself, either through the clearance or by breaking the barb off the barb-leg.

I have several times opened the stomach of a palu, but it was invariably quite empty. The natives, however, state that they frequently find feke (the squid) and ngufeke (another of the cephalopoda) in it. Now if the squid at great depths frequents holes in the vertical face of the reef, as it does in the surface-reefs, it is reasonable to suppose that a palu, cruising in search of food, would clamp its jaws on the protruding tentacles and attempt to pull the squid out of the crevice. I mean to suggest that the palu may - 29 thus have formed the peculiar habit of seizing and attempting to carry away its prey against an unidirectional resistance; a habit which alone, in my opinion, renders it possible for it to be caught in the manner in which it is commonly caught on this highly-specialized hook.

Vaitupu is a low-lying coral atoll, somewhat oval in shape, of an extreme length of about four miles, and containing two lagoons, one of which is the perfect wholly-enclosed “blue lagoon” of South Sea romance. As regards the physical formation, however, we are not here concerned with the interior but rather with the exterior aspect of the island. Beyond the circumferential belt of coconut-palms is a flat tidal reef varying in width from approximately 300 yards to half-a-mile. Except in a few places where there are submerged sloping shelves of no great width, the face of the reef slopes sharply to the hundred fathom line; then, as one proceeds out to sea, the soundings become rapidly deeper until, at an average distance of about one mile from the edge of the reef, there is “no bottom,” or at least none from the point of view of a fisherman who is not particularly desirous of paying out and hauling in continuously more than 250 fathoms of line, to say nothing of a 3-5lb. sinker. To qualify this general picture, one has to mention that there are shallower patches in places even farther out, and a particularly long, sloping shelf of submerged reef protruding from the north end of the island.

Now palu frequent the depths 80-250 fathoms (and perhaps greater depths, but I know of no one who has had the energy to operate more than 250 fathoms of line) and are caught, not on the bottom, but off from the sheer face of the reef. I have proved this many times by noting that at a successful fishing ground, fishing, say, at about 150 fathoms, the normal palu depth at Vaitupu, if the canoe is allowed to drift inshore with the wind or current, as the case may be, the line quickly becomes fouled on the coral, although, when payed out, it may have been nowhere near bottom. An experienced tautai (fishing-canoe captain), by taking bearings from land marks ashore, can tell, to within a few yards, the location of any of these tō (fishing grounds). Roughly speaking, there is a tō on either side of any reef - 30 having a sheer face or steep slope through the depths frequented by palu. The side of the reef or of the island on which palu are to be found on a certain night is determined by the direction of the weather. And an interesting point here is that a tautai will not usually go palu-fishing unless the weather has been in the same direction for at least three days. He then goes to the lee side of any tō (fishing-ground) he may select. This does not necessarily mean the lee side of the island, although that is usually preferred on account of the calmer condition of the surface.

There are conflicting opinions among native palu-fishers as to the merits of the period of the dark moon. Generally speaking this period is preferred for all kinds of night fishing, but palu are certainly to be caught on bright moonlight nights. Perhaps the preference for the dark moon is owing to the fact that flying-fish, the usual bait for palu, are taken in quantity by the torch-light netting parties only during that period. This, however, is by no means a sufficient reason, since flying-fish are often taken at sunset, irrespective of the moon, by being herded up into pockets in the reef.

Favourable judgment having been passed on the indications of sky and sea, four of the crew of the family-canoe, (including the tautai), assemble at dusk on the village beach. Here, there is a considerable amount of activity, some canoes returning with their catches from the sunset, flying-fish-netting parties; some, if there be no moon, preparing to set out with their coconut-leaf torches and long hand-nets, also after flying-fish, while one or two others perhaps, like our friends, are preparing their equipment for a night's palu fishing.

The tautai brings a large coconut basket containing two or three coils of thick line, averaging 80 fathoms per coil, his total supply of hooks, maybe five or more, since it is not uncommon to lose several by the attacks of sharks on the hooked palu, one or two sinkers of stone or scrap metal, and lastly a club (siki) 3 or 4 feet long and about 1½ ins. thick. One other member of the crew, the bow paddler, may be similarly equipped. All four have paddles, and the wooden shovel-shaped bailer is already in the bottom of the canoe.

When all is ready the tautai gives the word, the coconut-leaf covering is stripped from the body of the canoe, paddles - 31 and equipment are put aboard and the canoe carried down the slope of the beach to the water's edge; or, if it be low tide, it may be necessary to carry it about a quarter-mile, almost to the edge of the reef. Six or eight flying-fish for bait will have been obtained already from some relative returning from the netting parties, or the palu canoe may itself carry a hand-net and a few torches to collect its own bait on the way to the tō (fishing ground) selected. Having shot their canoe over the line of breakers at the edge of the reef (an adventure in itself) and paddled out to the tō, the tautai and the person in the bow set about baiting their hooks while the others hold the canoe in position against wind or ocean current.

To bait the hook, the scales are removed from a flying-fish and the body is split lengthwise. Head, tail and vertebral column are removed and one of the remaining sides is bound tightly, skin down, onto the barb and barb-leg, outside the knuckle, so that it extends almost from the tip of the barb to the lowest turn of the scarf-joint lashing. A smaller strip of the flying-fish is then bound transversely across the knuckle of the barb so that its ends hang free and have a flapping motion when in the water. These ends are said to resemble the tentacles of a squid.

The snood of the hook is then knotted to the end of the line and the sinker is fastened directly to the line, further up, so that it hangs less than a fathom above the hook.

Now-a-days rough stone sinkers are carried only in case of emergency. The sinker of former times is described as a somewhat elongated and rounded stone covered by a binding of kafa (thick, three-plait sennit braid) by which it was tied closely to the line.

Sinker and hook are now cast overboard and the line is allowed to run out quickly for about 80 fathoms; it is here held for a minute or two in case the palu are ranging higher than usual, and then, if nothing occurs, different depths are tested until about 150 fathoms of line have been payed out. If there is still no bite, it is usual to let the line run till the sinker strikes bottom (if within 250 fathoms) and then haul in again 5 or 10 fathoms. The line is then held at this depth for from 10 to 20 minutes.

- 32

The other line is similarly operated, but it is usual for one fisher to question the other as to the amount of line he has payed out and the two hooks are kept at different depths until it is discovered at what level the palu are biting.

If a tō (fishing ground) thus tested, is found unfavourable, lines are hauled in and the canoe moved to another tō, perhaps miles distant. This is then tested in the same manner, and if necessary, others, until one is found where the palu are biting well.

A length of 150 fathoms of thick line hanging vertically over the side of the canoe, has so much inertia that the sinker does not move up and down in synchrony with the rise and fall of the canoe on the surface swell, but lags considerably. The result is that the line is sometimes slack and sometimes exerts a pull on the hand equal to more than twice the weight of the sinker. Again, the struggles of a palu to take the bait are, for the same reason, not transmitted to the hand as a series of short sharp movements, but rather as a distinct downward pull, sometimes so strong as to be unmistakable, but in the case of a small palu easily mistakable for the natural pull of the line. It is thus difficult for the fisher to determine whether or not a palu is at his hook.

If he decides that it is a palu, he takes a steady strain on the line, being careful to raise his arm as the canoe slips down into the trough of the swell, and to lower it again as the canoe rises on the crest, thus keeping an even strain on the hook by compensating for the movement of the canoe. As soon as he feels a sudden release of weight, he hauls in as quickly as possible, and keeps on hauling or taking the strain during renewed efforts of the fish to swim downward.

When the palu is brought to the surface, it threshes around violently. It is now the part of a good fisherman to shorten his line in his right hand and slip the left hand out under the lower jaw of the palu seizing the barb-leg of the hook, near the base; no easy task with palu measuring more than 4 feet in length. (Some fishermen clasp the barb in the fingers and grip the shank-leg with the thumb). The head of the palu is then brought alongside the canoe and held, skull outwards, while one of the crew clubs it to death with the siki (the stick brought for this purpose. See Figs. 9, 16).

- iii - iv - v - vi - 33

Figs. 11, 12, show clearly the position of the hook in the mouth as the fish is brought to the surface. Sometimes it is found that the barb has torn its way through the floor of the mouth, and sometimes, that it has passed out through the gill opening and doubled round the lower jaw clamping both the jaw and the tissues of the floor of the mouth into the crotch between the shank and barb-legs of the hook. The figures also show clearly the reason for the attempt to give the inside of the crotch of the hook a U- rather than a V-shape. In Fig. 11 the hook shown is No. 2 of the first hooks described in this paper, and in Fig. 12 the hook shown is No. 3 of the same series. If reference is made to these two hooks, it will be seen that, although the maximum inside width of No. 3 is greater than that of No. 2, yet the interior of the U-shaped crotch of No. 2 is much wider than that of the V-shaped No. 3. Hence No. 2 hook has been able to encircle the jaw bone and is in little danger of breaking since the strain of the palu is in line with the shank-leg. No. 3 hook, on the other hand, in Fig 12, shows the strain applied to the lower part of the barb-leg, since the space in the crotch was not sufficient to enable the flattened bone of the lower jaw to twist round in it. The barb-leg of this hook is therefore in danger of breaking.

When the fish ceases to thresh about, the hook is removed and the body is lifted inboard and thrown in the bottom of the canoe.

It has been remarked by one writer that Ruvettus is phosphorescent. 5 In all the fish I have taken at Vaitupu, I have seen no phosphorescence that is an inherent property of the fish. Certainly there is an abundance of phosphorescence in the water on many nights. Sometimes it adheres to the line and remains bright for some minutes when the wet line lies coiled in the bottom of the canoe. Anything moving in the water stirs up quite a glare, and the palu can thus be seen for three or four fathoms before it reaches the surface.

No account of palu-fishing at Vaitupu would be complete without some mention of the fisherman's tapatapa. This is something in the nature of a semi-querulous banter chanted - 34 to the fish. It begins when the first line is thrown out and continues at random until the morning star calls the end of the night's work. The word-phrases of the tapatapa are very ancient and but imperfectly understood by the modern generation. This then, is the chant to the palu, with a free translation.

When one of the fishermen feels a bite, he says to the other, “Te palu ka fano ki te lou.” (Lit. “The palu is going to the hook.”) The other replies, “Sanga tonu ke mate.” - 35 (“Look to it that you get him.”) Then the first, “Tuku ke oko.” (“Wait a bit till he's on.”) Then, a moment later, the second will ask, “Ea, ko mate?” (“How now, have you got him?”). And the reply, “Ko mate.” (“He's hooked.”)

These words rarely vary, and are part of the conventional rites of palu-fishing.

On some nights sharks are very troublesome. Their favourite depth for attacking the hooked palu seems to be about 50 fathoms. Thus it frequently occurs that one labours, perhaps from 10 to 20 minutes hauling up a 5-foot palu through 100 fathoms only to provide a meal for one of these prowlers. Usually he leaves the head on the hook, but sometimes returning for this last morsel, there is a case of “the biter bit” and one has the compensating satisfaction of adding his carcase (malodorous though it be) to the collection in the bottom of the canoe.

When the fetu ao (morning star) rises over the dark distant line of the palm-jungle it is time to haul in the lines and make for the village; and on mornings when the tide is high, a very pleasant conclusion to the night's fishing is the exhilarating, swift glide of the canoe, on a single wave, from the boiling white surf-line of the outer reef right up to the white sands of the village beach.

The word palu connotes many kinds of deep-sea fish, only two of which are true Ruvettus. The Native classification is:—

Of the palu valevale (strange kinds of palu) two are illustrated in Fig. 17. The large fish is palu malau, and the small one palu kuakua, another name for which is palu ko mulo (scaleless palu). The colouring of palu malau is deep pink merging into white, while that of palu kuakua is irridescent-blue on the back and silvery-white on the belly.

- 36

The palu of this class take their names from other species of deep-sea fish which are caught on lines much nearer the surface. They are said to resemble both the palu and the fish from which they take their names. For instance, mangō is the native word for shark. Palu mangō is a shark resembling a “hammer-head” but it has certain points in common with the palu, one of which is the depth at which it is caught. Palu fai is a “ray” caught at the same depth. All of this class of palu which I have seen have the huge eye of the true palu, reaching, sometimes, a diameter of two inches; and all are caught on the native palu-hook, baited for palu.

Of the palu maoni (true palu), all the photographs of Ruvettus in this paper illustrate the palu fala species. This fish is of a reddish-brown tint, sometimes with large asymmetrical patches of a milky-white colour on the sides. The scales have peculiar little cleft horny excrescences which make the skin extremely rough to the touch. This palu attains an extreme length of about six feet and a maximum weight of about 60lbs.

The lavenga is much lighter in colour, in general, larger, and otherwise of the same appearance as the palu fala, but is comparatively rare. During four years' residence on Vai-tupu, I have seen only one, which was about the size of a normal full-grown palu fala, and I greatly regret that I neglected to examine it in detail. The natives say that it has characteristic differences in the shape of the head. They claim to have seen one a fathom and two hands'-breadths long.

Louis Becke writing 8 of Ruvettus fishing in the Ellice Group says of the fish, “The largest I have ever seen was nearly six feet in length, weight about 130lbs., and was caught off Oaitupu (Vaitupu) Island …”

Now, of all the fish I have weighed and measured, the largest was 63.5 ins. in length and weighed 46lbs., and in all, the weight has been almost proportional to the length. Of 24 palu fala taken within the last fortnight while the paper was being written, the smallest was 35 ins. in length and weighed 8lbs., while the average weight was 27.9lbs.

- 37

The older natives say that they have taken rare palu which had a greater than average girth, but it is such a far call from 46 to 130lbs. that one is inclined to think that Becke was taking advantage of the time-honoured fisherman's licence.

It would perhaps seem a graceless act to criticise the statements of a graceful and interesting writer of fiction, but, with the exception of Augustin Kraemer, no one else but Becke seems to have written anything about the palu and palu-fishing of the Pacific, and it is necessary therefore to draw attention to several inaccuracies in Becke's writings.

Waite, who identified the Funafuti palu as Ruvettus pretiosus, quotes Becke, “a full-grown palu would weigh up to 150lbs….. the jaws are toothless …. The average size is about three or four feet, and weight 40 to 60lbs.” 9

The probable maximum weight has already been indicated. The average length of palu taken here is 52.6 ins., and, as previously stated, the average weight 27.9lbs. The usual weight of a 3-foot palu does not exceed 10lbs., while a 4-foot specimen may reach a weight of 30lbs.

Again, in the same Memoir, Becke is quoted, “It is prized above all other fish in the Line and Ellice Groups.” This is certainly not so. Speaking, not only of Vaitupu, but from a general acquaintance with both the Gilbert (Line) and Ellice Groups during the past six years, I can definitely state that of the large edible fish palu is one of the least commonly sought. The majority of the people are generally ashamed of its pronounced laxative effect and are therefore afraid to eat it in the quantity which they usually require to make a satisfactory meal. (Witness! Te ika fakama vale tagāta mo fāfine—“the fish which inevitably makes both men and women ashamed” already quoted as part of the tapatapa chant.)

In another place 10 Becke says, speaking of palu which he caught at Nanumea, in the Ellice Group, “Although he (the palu) makes an occasional spurt, he grows weaker and weaker as he is dragged toward the surface and, when lifted into the canoe, is apparently lifeless, his large eyes literally - 38 standing out of his head, and his stomach distended like a balloon. So enormous is the distension of the bladder that sometimes it will protrude from the mouth, and then burst with a noise like a pistol shot.”

I have never seen bulging eyes, distended stomach or protruding air-bladder. When the fish is brought up, the eyes are full and, I should say, normal, in contradistinction to their withered and somewhat sunken appearance the next day. The stomach is almost always empty and the sides consequently loose and flabby. Native fishermen have informed me that the swim-bladder of the palu malau (see Fig. 17) always protrudes and sometimes bursts when the fish is brought in. They state that they have never seen the bladder of Ruvettus protruding.

Later, (loc. cit.) Becke claims to have caught, at Beru, in the Gilberts, a palu measuring 6 ft. 10 ins. which, although in rather poor condition weighed 136lbs!

Coming now to the effect of the oil upon the partaker of the flesh, I will merely note that my personal experience corroborates all that Kraemer has written. 11 The effect of the oil is identical with that of our medicinal mineral “lubricants.” It is the fact that its victim is usually unaware for some time of the effect it has produced that has given rise to the expressions “tragikomisch” and “fakama vale” previously quoted in this paper.

The flesh is soft and palatable, but, even after boiling, it retains a considerable quantity of oil and one cannot safely eat more than about 4 ozs. Salted, sun-dried, and smoke-cured, it is delicious.

The natives cook it in the ordinary culinary oven made of small coral stones heaped over the glowing embers of the fire. The flesh is wrapped in plaited, green coconut-leaflets which allow the oil to drain away freely as it is being cooked.

- 39

EASILY the most exhilarating sport in the Ellice Group is the hunting of bonito. It is the ambition of almost every young man to become an accomplished tautai (fishing captain) and to have the honour, accorded the most skilful bonito-fisher on a ceremonial expedition, of leading the canoe fleet ashore through the surf.

The flesh of the fish is highly relished, both cooked and raw, and an abundance of bonito is usually a signal for communal feasting.

The following brief description of bonito-fishing on Vaitupu Island will commence with a description of the apparatus used. The second section will be devoted to the method of fishing, and the final section will give an outline of the procedure of a ceremonial expedition.

The apparatus consists of a long rod (kofe) to which is attached, each on a separate line (afo) from two to six (the usual number is four) pearl-shell lures (pa). A description of such parts of the fishing-canoe as may truly be considered bonito-fishing apparatus may be found in the chapter on Canoes.

(a) Te kofe. The rod is usually between twelve and eighteen feet in length. It consists of two separate pieces, the proportions and general shape of which may best be illustrated by diagram (See Figs. 19 and 20). The butt (te tu kau) is usually given a length of from eighteen inches to three feet, varying according to the fancy of the individual owner.

The long outer extremity is made of a slim bough of milo (Thespasia populnea [?]) fetau (Callophyllum inophyllum) or puka (Hernandia peltata [?]). Imported bamboo, when obtainable, is preferred to the native woods. The butt is made of mature milo wood or of kanava (Cordia subcordata [?]).

- 40 - 41 - 42

In the neighbourhood of the butt-joint a piece of old fish-netting is wrapped around the rod. Into this the barbs of the lures, on their taut lines (afo), are hooked when not in use. This netting is known as saf enga a pa or fakanatinga a pa. In former times, it is said, the fakanatinga was specially made for the rod (See Fig. 22).

(b) Te pa. The pearl-shell lure. The principal parts of the lure are:—

Lures vary in length roughly from two to five inches. The general proportions of the shell-piece (te pa) may be seen from the sketches in Fig. 18. Modern pa are cut from the pearl-shell with hack-saws and ground to shape on a coral grinding-stone (punga). The pearl-shell is cut so that the thin outer edge becomes the barb-end (te ulu) and the thick base the point of final attachment of the line (te mulimoa).

Most pa have the ulu-end cut square. Some, however, have a rounded indentation (e fakamanga te ulu) (Fig. 18c and d). And others have small lateral projections at the extreme end (Fig. 18f). There seem to be no reasons for these variations except the momentary whims of the makers (tautai—fishing captains).

The width varies as the length, roughly from ¼ to ½ in., and the thickness depends first on the size of the original pearl-shell and secondly upon the amount of grinding given the exterior surface to bring out the required colour. The side edges, especially towards the mulimoa, taper inwards (considering the convex surface the exterior).

The colour or combination of colours and sheen which may be obtained by proper cutting and grinding is one of the most important considerations in the manufacture of pa. The exterior of the base-end (te mulimoa) of the shell-strip, when smoothed, invariably comes up white, but the lip-end (te ulu) may opalesce with any number of colours, the intensity of which may be regulated by the amount of grinding.

- 43

Four different kinds of pa are distinguished on Vaitupu and a tautai (fishing-captain) always endeavours to have available at least one of each kind. The distinction is one of colouring only.

The commonest is (1) te ngu. This is pearl-grey or pearl-white all over. It takes its name from the colour of te ngufeke (a cephalopod) on which bonito are said to feed.

Of the iridescent kinds, the most usual is (2) te fululupe. This takes its name from the iridescent sheen on a bush-pigeon's wing-feathers, which it is said to resemble.

The others are (3) te nanu-talakisi, a reddish colouring said to resemble the colour of a small fish (te talakisi) and (4) te laumilo, a distinct opalescent-green. It takes its name from the bright-green leaf of the milo tree. Sometimes dead white spots are to be seen on the exterior of a shell. A lure having these is referred to as sungalu.

In ancient times a pearl-shell lure having a desirable combination of all the different colours was referred to as lanusamasama. The modern equivalent seems to be the fulu-lupe mentioned above.

The reason for these different colour-combinations and the necessity for having a specimen of each may be mentioned here. It appears that the bonito has a predilection for variation of diet, and whereas he may be found attacking the shoals of one species of small fish to-day, next week he will probably be found only among those of another species. It will have been noticed by the reader, and the makers themselves claim, that the pearl-shell lure is shaped on the lines of a small fish. (An apparent anomaly is, however, the naming of the fore-end te muli-moa and the after-end te ulu).

The fish on which bonito feed are:—

• (a) o, a small fish, silver and bluish-green in colour.
• (b) o tala, similar to te o.
• (c) kavaliki, a greenish fish attaining a length of about 3 in.
• (d) ngu, cephalopoda.
• (e) kaki, a small, greyish fish about two Inches in length, with protruding eyes and a bulbous head.
• (f) kalo, a greyish fish attaining a length of from 2 to 3 ins.
• (g) Small soft crustaceans less than 1 in. in length.

It is for shoals of o that the laumilo lure is said to be especially suitable, but any one of the coloured lures may be used when the bonito are feeding on any of the first three - 44 fish listed above. A man working in a shoal from which the canoes are lifting bonito will change from one lure to another until he finds the colour that the fish take most readily that day. The ngu seems to be useful at any time, but particularly so when the bonito are feeding on the four last-mentioned foods.

I have heard it stated by a tautai from Nanumea Island that the ngu are particularly useful at daybreak, and the darker colours in full daylight.

The barb may be made of hawks-bill tortoise-shell (te una), the lower jaw-bone or the cranium of a pig, whalebone (ivi o taf ola), or whale's tooth (nif o o taf ola). Of these, the last-named is the most highly prized, and a barb made of it is called te lei. The majority of barbs nowadays are made of tortoise-shell but, since the hawk's-bill variety is extremely rare in the group most of this has to be imported. I have heard it said that, in pre-European times, hard thick coconut-shell was occasionally used.

The barbs are of an average thickness of about in., and vary but little in size and proportion from that illustrated in Fig. 21. The shank-piece of the barb on some lures has three holes pierced for the lashings and line-attachment, but on others it has but two.

The line (af o) was in former days made of young Hibiscus bark (lau fou). This is still sometimes used, but has been almost displaced by cotton line. It is usually about four fathoms in length.

The lashing-line (te alaala)) was also formerly of lau fou but is now usually made of strong sewing-cotton. It is in three pieces—(1) lashing the base of the barb-leg and the feathers to the end of the shell; (2) lashing the barb-leg to the shell; (3) seizing and protecting the main line between the barb-leg and the muli-moa, and forming an auxiliary connection between the barb-leg and the lashing-hole in the muli-moa.

The feathers are almost always the short breast-feathers of any white sea-bird. Four or six feathers are arranged in a double-ended bunch laid transversely under the end of the shell and attached to the barb-lashing in that place.

The barb-wedges are made of the stems of two coconut-leaf pinnules. They are driven under the barb-lashings on - 45 each side in the angle between the barb and the interior surface of the shell. They serve to give the barb a vertical stability besides tightening the barb-lashings.

The operation of lashing together the various parts of the lure was in former times surrounded by a great deal of ceremonial tapu, some of which persists in modern times. The tautai (fishing-captain and maker of all important pieces of his own apparatus) would sit on a mat in his own house surrounded by all the un-assembled parts of his lures, i.e. the shells ground to shape and pierced, the barbs cut and each shaped to fit the surface of the shell for which it was destined, af o lines, a supply of alaala, a supply of feathers, and a number of coconut-leaf-pinnule stems.

No female or child was allowed to approach him; neither might any person be present other than the adult males of his own family.

The first operation was the barb-lashing (founga o te manga). Holding the shell concave surface upward, the barb was set in place. Then with the loop of a piece of alaala (lashing-line) hanging below the shell, the two ends were threaded through the lashing-hole in the barb-leg. The ends were drawn up tight and crossed underneath. This process continued until there were (usually) six complete turns. The ends, having been crossed on the exterior surface of the pa after the final turn, were drawn up, one on each side, and knotted twice around the lashing-turns in the angle between the barb and the shell.

In ancient times, it is said, the tautai must tie all knots on the left-hand side of the lure with his left hand, and those on the right with his right hand.

The lashing of the base of the barb-leg was left in order to be done in conjunction with that of the feathers (senga) to be described later.

The next operation was the making of the auxiliary connection (alaala-loa) between the barb-leg and the hole pierced in the muli-moa of the shell (founga o te fatafata a pa—lashing of the breast of the lure). It should be noted that this auxiliary connection is not present on all lures. Those which have it are said to be fou tonu (properly lashed) while those without are said to be fou talili (just tied up).

An end of the alaala line was threaded through the hole in the barb-leg nearest to the muli-moa. A loop was - 46 then made with a slip-knot and drawn up tight on the barb-leg. The other end of the alaala-line was then threaded through the hole in the muli-moa from the right-hand side, (i.e. with the lure held so that the muli-moa was directed towards the maker). It was drawn up tight and again passed through the barb-leg from the right-hand side. This operation was continued to make two or three complete loops between the muli-moa and the barb-leg. The holes were always threaded from the right and so the lines became crossed immediately in front of the barb-leg. The free end was now used to seize, in two single knots, immediately in front of the barb-leg, all these lashing loops.

The main line (afo) was then inserted through the same hole in the barb-leg and drawn through sufficiently far to double back on itself as far as a point about in. beyond the hole in the muli-moa. Some tautai make a single knot with this line on the barb-leg before doubling back the free end. Others, after drawing through the free end, open a strand (lino) of the long end just in front of the barb-leg and thread the free end through this.

The long end of the line was now twisted (pekai) four times round the great toe of the right foot. The exact number of times was part of the very deliberate ceremonial carried out by the tautai. The methods of tautai in another family group might differ in such details.

The lashing was now continued with the free end of the alaala-line which was used to seize in four turns the doubled afo-line which lay uppermost with the loops previously described underneath.

These seizing turns vary considerably in nature and number among the different families and on different islands. They are held to be a most important point in the assembly of the lure. Generally speaking, they are always in a single flat layer extending from the barb-leg from ⅛ to ½ in. along the afo-lines.

I presume, although tautai do not say so, that this close layer of seizing turns is meant to protect the line from the teeth of the bonito. It is interesting also to observe that tautai, blindly following hereditary custom, seem quite unaware of the obvious reason for various devices. The alaalaloa loops, for example, which obviously form what I have called them, an auxiliary lashing to save the barb and the - 47 fish in the event of the main line (afo) being cut by the teeth, are looked upon by most tautai as a necromantic device to bring luck to the fisher.

When the prescribed number of seizing turns had been put on, the free end of the alaala-line was taken back to the muli-moa on the left side (muli-moa was now directed away from the maker and was held close under the afo-line fastened to his great toe), and, the slack being held on that side, it was threaded through the hole in the muli-moa from the opposite side and then brought over tightly to seize the double afo and its own slack immediately above the hole. This part of the lashing was called (tuituinga o te muli-moa). By the tautai whose method forms the main part of this description, this tuituinga-lashing was always given four very tight turns.

It should here be noted that all the longitudinal lines between the barb and the muli-moa (founga o te fatafata a pa) were left sufficiently slack to allow of their touching the surface of the shell on pressure being applied from above. This, it was claimed, lessened the danger of the line being severed by the fish's teeth. I have seen lures from Nanumea Island, however, in which the founga o te fatafata a pa was quite taut; but in these, the protective seizing turns immediately in front of the barb were much more numerous and more efficiently applied than is the case with Vaitupu Island lures.

The free end of the alaala-loa was now taken four times transversely with respect to the vertical tuituinga-turns just described, seizing these between the afo-line and the upper ridge of the muli-moa.

It was then given four figure-of-eight turns, seizing alternately the doubled afo-lines on either side of the vertical tuituinga-turns. The free end was now knotted off with a double half-hitch on the afo behind (i.e. away from the lure) the muli-moa.

This part of the operation may now be considered complete, or the tautai may, at his discretion, reduplicate the figure-of-eight turns in order to stiffen the assembly at this point. It will be noticed that the figure-of-eight turns have the effect of making a kink in the afo-line. This assists in taking the strain off the shell, for it operates in a manner similar to that of a bell-angle throwing the strain directly - 48 from the barb on to the main afo-line exterior to the hook. On some Nanumea lures examined during the preparation of this paper, the kink mentioned above was not observed.

Lastly, the free end is brought from the exterior knot to the doubled afo just inside the muli-moa ridge, and is there knotted tightly in a spaced series of from two to four straight single knots, after which it is cut off.

The lashing of the feather-lure (founga o te senga) was the next operation. A piece of alaala-line about 1 ft. in length was now threaded through the hole in the base of the barb (i.e. at the ulu-end of the lure). The ends were left hanging free while another coil of four turns of alaala, about 1 in. in diameter, was placed lengthwise under the shell in this region. The free ends of the first piece, catching up this coil, were used to lash the base of the barb in a manner exactly similar to that already described for the barb-lashing (founga o te manga), except that the free ends, after seizing the lashing-turns in a double knot on either side of the barb, are not cut off, but are brought round to the base of the barb where they are joined in two single knots spaced by about ¼ in. They thus form a tail, about ½ in. long, which goes by the same name as the feather-lure (senga). On Nanumea Island the lashing-line tails on the lure are known as vae-kaviki (crabs' legs).

The double-ended bunch of feathers is now passed through the coil which was left hanging below the last-mentioned lashing. The concave face of the bunch is usually directed downwards. The four loops of the coil are then drawn up tight on the stems of the feathers and the free ends, being knotted, are also given a senga (tail) similar to that described above.

The senga then consists of two tails of alaala-line projecting about ½ in. beyond the ulu-end of the lure, and a bunch of feathers projecting about 1 in. on either side.

The final operation is the insertion of the barb wedges (seketi). The fine needle-like midrib of a dead coconut-leaf pinnule is now forced under both the barb-lashings on either side of the barb. These are broken off flush with the base of the barb and with the inner end of the barb-leg. In cases where the barb is thick and has a good seat on the shell, the seketi may be omitted.

- 49

When the weather is favourable and cloud and lightning omens have been indicating an abundance of bonito, there are few canoes of the village fleet which are not on the open sea before day-break. Most of these scout along the leeward shore, but operations are by no means confined to that shore. Fleets of two and three canoes may frequently be seen almost out of sight of land or bobbing about on the weather side, the sport of the unceasing trade-wind.

The presence of a shoal of any of the small fish on which the bonito prey is almost invariably indicated by a whirling black vortex of sea-birds poised over the fish and individually falling and rising vertically as fast as they can devour the fish which they secure. There can usually be distinguished in such a swarm te ngongo, (noddy—Anous stolidus Linn., and white-capped noddy—Micranous leuco-capillus Gould, te matapula (white tern—Gygis alba Sparrm.), and occasionally te katafa (or upuitoe) (greater frigate-bird—Fregata aquila Linn.). The arrival of the bonito is indicated by their leaping and also by sharp splashes, as of a bullet striking the water immediately beneath the flock of birds. The canoes, sighting a flock of birds gyrating in the manner indicated, paddle swiftly toward it, but as a rule await an indication of the presence of bonito before dashing in.

On approaching the shoal, the tautai (fishing-captain) of each canoe lifts his rod out of the rests on the outrigger, detaches one of the four lures on the rod from the safenga a pa, and, casting it out, sets the butt-end of the rod in the matapili (see chapter on canoes) and lays the rod across the tokoulu on the stern-cover. The tautai sits so that his buttock is in a position to feel any movement of the rod caused by a bite.

On feeling a bite, or on hearing a shout from one of the canoes behind him, he throws his paddle forward with his left hand to the ta fa (paddler who occupies the seat on the after boom) and with his right hand works the rod, pivoted in the matapili, so as to draw the fish steadily out of the water. Then, standing, he makes it swing with considerable force against his abdomen. Here he catches it - 50 under his left forearm, caudal fin to the left and with his left thumb, pressing the interior face of the lure, disengages the barb, the right hand being still occupied with the rod. Then simultaneously, the fish is dropped into the canoe at his feet, the lure is cast back into the sea and, except in a shoal, the rod is again leaned over the tokoulu. During the operation of swinging in the fish from a sitting position, some tautai obtain a purchase on the side of the canoe by setting the toes of the right foot against the calf of the left leg.

If it happens that the bonito caught is a stray one, the tautai will sit again and go on paddling. Here it should be noted that in the proximity of bonito a tautai never paddles on the port side of the canoe lest a bite should find him unprepared.

In a shoal the canoe is brought to a standstill, the tautai remains standing and trolls his lure lightly over the surface.

If the fish are plentiful and are biting well, the lure may not even touch the water but may skim along just above the surface. This practice makes it possible to catch a much greater number of fish in a given time, for the fish is drawn clear of the water almost as soon as he bites, and the time usually lost during the struggle that ensues when he is able to get his head down, is thereby saved. Working in this manner, it is no uncommon thing for a single tautai to take as many as fifty fish out of one shoal.

It may happen, when a canoe is going ahead, that a pala (kingfish [?] of an average length of 4 ft. 6 ins.) will take the lure. This fish jumps clear out of the water, and the tautai having felt the bite and swung round to smack his right hand on the rod and swing in the fish, seeing the pala in the air, must shout quickly, “Ou nifo!” (“Your teeth!”) This is said to have the virtue of causing the pala to become unhooked without cutting the line and breaking the barb off, which accident usually happens in such a case. After a line has been thus cut by a pala, it is said to be unwise to fasten another lure to it, as the new one would be likely to suffer the same fate. The ill-luck attaching to such a line might, however, be dispelled by dipping the line in water which had been heated in a coconut shell.

- vii - viii - ix - x - 51

There are two kinds of communal fishing, one in which the two sides of the village (ngutu malae) compete, and the other in which the whole village-fleet goes out under one command. The first is called faiva and the second tau-tau. Either kind is commonly followed by a communal feast.

On the day before such a ceremonial expedition after bonito, all the tautai concerned spend many anxious hours attending to the minutest details of their bonito-fishing apparatus. Lures are taken to pieces and re-assembled. The rod is examined. The canoe crew (faoa) is selected and arranged. The sky is subjected to a continual scrutiny for propitious bonito-omens, and its changing moods are reflected on the faces of the tautai. In fact, of all their various activities on land and sea, I know of no other which creates in the heart of the Ellice Islander one half the enthusiasm, careful preparation, and willing obedience to the tapu (abstinence) regulations, that is occasioned by the declaration of a ceremonial bonito-expedition. At night the tautai lies by himself. He rarely sleeps, but remains in a kind of doze from which he wakes at frequent intervals. On each such awakening he proceeds to the sea and bathes before returning to his bed. The explanation given is that sleep is impossible to an earnest tautai on account of the excitement inspired by the occasion, and that the bathing helps to brighten his eyes for the morrow's work.

About one hour before dawn the tautai proceeds to the sea to bathe, this time wetting all his garments. He then goes to his own canoe with all his apparatus. After uncovering the canoe, he fixes his four lines (afo) to the end of his bonito rod. The barb of each lure is inserted first in the safenga a pa (see above) and the line is drawn taut to the end of the rod where it is fastened in a simple double half-hitch, the free end is then secured along the rod in a series of single hitches. With many tautai the number of hitches given the free end of each line depends upon the “lucky” number of his lures. It was noticed in one case, for example, that a tautai gave four seizing turns to the alaala-loa seizing near the barb, four tuituinga-lashing turns on the muli-moa, and four single hitches as a finish for the alaala-loa - 52 lashing. He gave this as a reason for employing four single hitches evenly spaced on the rod as a means of securing the free end after the line had been fastened to the tip of the rod. With other tautai the “lucky” number is six, which usually corresponds with the number of complete turns taken in each of the main lashings on the lure. It is tapu for anyone to approach while the tautai is thus fastening a lure to his rod (tau te pa ki te kofe).

He then sets the butt of the rod in the matapili and leans it against the tokoulu of the stern cover. On ordinary occasions he would simply lay the rod on the lango-kofe forks on the outrigger (See chapter on Canoes).

All these preparations being completed, the tautai goes to awaken and assemble his crew. The canoe is carried down to the sea, the word “All aboard” is given, and they paddle out to cross the surf-line and join the fleet of similarly prepared canoes which are now assembling beyond the reef-indentation which they call a passage (ava).

In ancient times, it is said, a tautai on reaching the muli-ava (i.e. the angle of the reef-indentation, for there is no true passage) must jump overboard where the water was about waist deep, catch up some seaweed or slime from the rocks and smear it over his breast, after which he would climb aboard once more and put his canoe through the surf. This necromantic rite was called tapune ki limu o tai, and is often referred to in old songs.

The following verse, from the beginning of an old bonito-hunting song, has some bearing on the points already made:—

Mangi te mana taiao	Morning clouds threaten rain
Ta olo o susue te vaka	(Nevertheless) we go to uncover the canoe
Mamulu i muli vaka	Jump over the stern (into the sea)
Tapune ki limu o tai.	Perform the necromantic rite of rubbing seaweed (on the breast).

There are always several experienced tautai who automatically lead and exercise authority over all maritime affairs. These are known as tau-moana. Now on a ceremonial expedition the senior tau-moana directs all operations. When the canoes have assembled outside the surf-line, he may lead them all off in one direction, or he may split them into two or more fleets to exploit several places at - 53 the same time. He, however, controls the whole fleet and will allow no tautai to act on his own initiative. Thus, if, after long paddling, the fishers become thirsty, he may raise his paddle, (the sign for all canoes to come together), and direct that each canoe send one or two of its crew to swim ashore through the breakers and climb for drinking nuts with which they must swim out again to the canoes.

The fishing itself is carried out as described in Section 2 above. The tendency to ritualize and the strictness with which breaches of etiquette are condemned and punished are, however, well illustrated by the following.

It is tapu for a tautai or any member of his crew, during the attack on a shoal, to drop overboard any article of fishing-apparatus or clothing or canoe equipment. If any tautai should be so clumsy as to drop a fish back into the sea, either off his hook or out of his hand when unhooking it, he is considered to have reached the lowest depths of shame and debasement. In ancient times, an offence of this nature would secure for the unlucky fisher a very heavy punishment. His canoe might be broken up, or he might be forbidden to fish for a whole season. In modern times the clumsy tautai is often required to pay a fine of a stipulated number of coconuts.

The reason for this regulation is said to be that anything falling in the water is likely to scare away the whole shoal of bonito. Especially is this the case when a bonito drops from the hand into the water, half stunned by the force with which he has been swung against the abdomen of the tautai, and swims straight downward with a sickly and very noticeable tail-spin motion, to be followed immediately by the whole shoal. Every native fisherman has a strong conviction that such an accident has a most unfavourable effect on the fishing, not only in the place where the accident occurred, but all round the coast of the island. They even go so far as to say that by some telepathic communication, at the very moment a shoal of bonito follows one of its number down into the depths, all shoals which may be attacking surface food in any part of the island go down also and do not re-appear that day.

In the thick of a large shoal, the excitement is indescribable. The canoes dart up at top speed, the ta-fa paddler in each dashing spray from his paddle over the tautai's lure - 54 where it is trolling in the water. There are cries of “Foe! Foe!” (“Hither! Hither!”). From the rear canoes can be heard, “Maté! Te vaka tela ko tautaloa!” (“Hooked! That canoe has its first bonito.”). The first bonito caught by each canoe goes by the name of tautaloa. The sea-birds, disturbed by the canoes, dart hither and thither with raucous cries. The water is dotted with the sudden sharp plunges of the leaping bonito, and in places the surface is churned by the ugly swirling bodies of huge green sharks. Over all is heard the mocking falsetto cry of the tautai working their rods to the limit of their skill, calling to the bonito “Tulina! Tulina!” (“Chase it up! Chase it up!”). And ever the noise becomes louder with the lusty slapping of hundreds of hard fresh-caught tails against the drum-like sides of the canoes until it is quite impossible for a man to hear his neighbour's remarks.

The sudden lull when the shoal departs and the dying fish cease to drum discovers the tyro with a racing pulse and the short, incomplete breathing of intense excitement. Then, as things become calmer, some of the canoes begin to scout about once more while others remain idle (palapala) congratulating each other on their luck, or commiserating their misfortune. The democratic spirit may show itself in cautions given even by diminutive juniors to the tautai of their canoe, “Mata ma kemo” (Lit. “lest you blink, ” i.e. “Keep your eyes open in future”). Someone will stand and search the horizon, and seeing the tautai rods of another fleet working in the distance will say “Te fua-vaka e tau kaseu.” (“The fleet is standing to it”). Then there is a race over to the working fleet and the game begins anew.

Bonito-fishing normally ceases about mid-day. On ordinary occasions unless the catch of bonito has been unusually good, many of the canoes make for the pala (king-fish) grounds and remain there enticing the fierce pala into the running noose until late afternoon. On a ceremonial expedition, however, the signal—the raised paddle of the tau-moana—is given to assemble as soon as it becomes evident that the bonito have stopped feeding on the surface shoals. The canoes gather at a point almost opposite the village. Here, the crew of each canoe counts carefully the total number of bonito caught by their tautai. The tau-moana then stands in his canoe and calls the name of each tautai, who immediately replies, stating the number of bonito in his canoe. - 55 When all have reported their catches, the total number of bonito for the morning's fishing is computed and announced by the tau-moana. Then all tautai are commanded to set their rods at the trolling angle in the stern (with all lures fast to the rod, however) and the names of the tautai who have made the best catches are called in order of precedence. These set off and shooting the breakers in the order given make for the village beach. After a short interval the word is given and, with rods swaying and paddles flashing in the sun, the remainder of the fleet dash for the surf-line by the village “passage.” Twice on the way they are called to a halt and each tautai standing and seizing his rod first elevates the tip and then lowers it into the sea (ta ki kofe) three times in a kind of grand salute.

Shooting the breakers with consummate skill, they glide in swiftly over the level reef in front of the seething rollers with great fountains of spray shooting up in front of their bows and drenching the laughing, shouting paddlers.

When near the beach, they are met by a rush of women (telekanga) bearing baskets of foodstuffs which they have prepared. This food is passed over to the men while the women seize and eat the remains of whatever number of bonito the tau-moana may have given permission to be cut and eaten raw (ota) while at sea. The women then carry all the fish ashore and throw them on some mats on the malae. Here a division is made among the various families for cooking purposes.

Meanwhile the men have bailed and dried out their canoes, carried them up to their stands (lango) on the beach, and recovered the hulls with their heaped pairs of coconut leaves (taomanga).

Later, the whole island assembles in the meeting house and feasts sumptuously on cooked bonito and various other kinds of prepared foods, named in general talitali. 12

- 56

IT is said that the eyes of the intelligence of the deep are the little formless wriggling masses of phosphorescence which are to be seen dotting the surface of the black depths on almost any dark night. These are known as pula o tai—the flowers, or (a not impossible translation), the open eyes of the sea. They are accustomed, it is held, to slip over the tidal reef with the tides in the daytime and keep an eye on the doings of all tautai (fishing-captains) on shore. If it should happen that they discover a tautai breaking one of the special fishing tapu (prohibitions), they report this to the bonito shoals and to the pala on their return to the deep and so arrange matters that the unfortunate tautai will be unable to catch any fish. 13

There persists to this day a shame-faced credulity in the truth of this ancient belief, and it is probably on account of this that the fishing-customs of former times have yielded least readily to the sophistication that emanates from modern Samoa. The strict observance of all tapu-regulations and fishing-etiquette is, therefore, just as necessary to the snaring of pala as it is to the lifting of bonito.

Pala-fishing is the complement to bonito-fishing. When the bonito cease feeding on the surface shoals, all canoes, unless the catch has been very heavy, proceed at once to the pala grounds. Here, the tautai of each cuts up a bonito. The head, with gills attached, is severed from the body and attached to a piece of line of medium thickness and about five fathoms in length (te uka o te ulu—the head-line). To the free end of the head-line is knotted a fine line of a length of fifteen to twenty fathoms (te uka taki—the trolling-line), and to the other end of this trolling-line a piece of bonito breast about three inches long by one inch wide is attached by means of a short piece of very fine, easily-broken

- xi

- xii Page is blank.

- 57

line. The reason for this is that, while paddling along, the tautai holds the trolling-line in his teeth, and so, if the line were not easily broken, a heavy pala taking the bait would probably cause him to suffer an elementary dental operation.

The canoe is paddled about, usually not more than a few hundred fathoms from the reef, until the tautai feels his bait taken. When this happens, he calls out immediately “Pala ka vau!”(“A pala is coming!”) The paddlers hold water with their paddles and bring the canoe to a standstill while the tautai, standing and facing the stern quickly coils in his trolling-line. In the meantime the tāfā (paddler who sits on the after-boom) stands also and commences to fling out the head of the bonito to the water on either side alternately in order to attract the pala up to the canoe. After the trolling-line has been coiled in, the tautai takes the headline from tāfā and goes on flinging it about to his own accompaniment of a fierce volley of banter (tapatapa) calculated to anger the pala and make him lose his discretion.

The whole operation of enticing and catching a pala is accompanied by this running fire of teasing remarks to the fish, uttered with a traditional semi-falsetto intonation. The conventional phrases of the tapatapa are divided into three periods. The first is called tapatapa taki—the trolling tapatapa, and is uttered by the tautai through his teeth clenched on the taki line while paddling around awaiting a bite. The second is named tapatapa ulu (the head tapatapa) and is called out while the head is being flung about from side to side. The third is employed while the canoe is idling (palapala i lunga).

The following are some examples of the tapatapa taki:—

• “Pala kai oso—e, liu ma-a-a-i.” (“Pāla who jumps to the attack, turn hither.”)
• “Muna a tautai i te pongipongi, pala e lua.” (“The canoe-captains said, just before dawn, that two pala would come along together.”)
• “Tai fua mai tena ngalu.” (“His wave [i.e. the wave he causes by his disturbance of the surface] is about to roll in.”)
• “E! pala—e! Te kili pulepule, liu ma-a-a-i.” (“Hey! Pala with the striped skin, turn hither.”)
• “Foe! foe!” (Hither! hither!”—looking back at the taki line.)

The following tapatapa taki are from other islands of the group:—

• “Te ika, ata foe.” (“Fish who dashes to the attack.” [?]).
• “Ai aue! te piso uli!” (“Ai aue! black-head!”).

- 58

• “Te ngutu kao!” (“Long-gape!”).
• “Taulua mata kitekite.” (“Two sharp-eyed [pala]”).
• “A! te namo!” (Ah! [in disgust] [This is only] a lagoon.”). 14
• “Kainga, seai ni tangata!” (“Family [dwelling-place] without people in it.” Said when there are no bites on a certain fishing ground.)
• “Aia, ka tuku ei?” (“Why do you abandon [the game of coming up to my bait?”]).

The tapatapa ulu consists of a series of short staccato sounds made up of two words, “ita” (anger) and “tu” (appear, or approach) with misplaced accents.

• “Itá, Itá, Itá!
• “Tú itá
• “Tú, tú, tú!”

The word tu is uttered with an extremely short vowel.

The following are the conventional phrases of the banter used on Vaitupu when the canoe is idling with the bait hanging deep.

• “Fakauta mai te amonga niu i te katea nei!” (“Let appear [to be brought into the canoe, as would the killed fish] a carrying-load [two bundles of ten] of coconuts [fig. for two pala], a-starboard here!”). 15
• “Pāla mata ita na!” (“Pāla there with the angry face!”).
• “Mafulifulina!” (“Twisting and turning one!”).
• “Pāla mo se tafito niu na!” (Pala there, large as the bole of a coconut palm.”)
• “Tu! tu! tu i te malu ke sao koe!” (Approach! approach! approach in the shade [of this cloud] that you may be safe.”)
• “Tu! tu! tu! Tu aka o tapa ake tou taumafa!” (“Approach! Approach! Approach! Come along and eat the food prepared for you.”)
• “Tau poa malie nei!” (“Your dainty bait here.”)
• “Fakalilo mai mua i tou lalo-peau nei.”(“Come warily neath the foam of your wave which is breaking here.”)
• “Tou laumatangi! Ko se tapa ai koe?”(“Your breezy way [of dashing about]! Why don't you stay still?”).
• “Tamaliki!” ([Bold as] a child.”)

To return to the fishing itself, while the tautai and the tāfā have been occupied with the trolling and bonito-head lines, the tino loto (person who occupies the seat on the boom immediately in front of tāfā), has been setting the ikupenga. - 59 This is a piece of moderately heavy line, 4½ fathoms in length, one end of which has an eyelet about one inch in diameter through which the line is doubled to form a running noose. The other end is tied to another heavier line about twenty fathoms in length. Lying on the outrigger platform (kaupalepale) with the noose line is a straight slender stick about 9 ft. long, (te kausele). The tino loto, on hearing the tautai shout “Pala ka vau, ” takes first the noose, and, inserting his thumb in the running line distends it to a length of 1½ fathoms. Then, holding the noose by the running-knot with his right hand, he twists the running line on to the tip of the kausele with two or three loose turns, after which, taking both the butt of the kausele and the running-knot in his right hand he passes the whole to the tāfā, who keeps it in readiness for the tautai.

Meanwhile, it is the duty of the bow-paddlers to keep the canoe into the wind and to prevent it from drifting.

The tautai throws small pieces of chopped-up bonito flesh to the port quarter of the canoe (muli-ama) in order to entice the pala to head in from that direction and so cross the stern from port to starboard.

Immediately the pala appears, he draws the bonito-head quickly into the canoe, seizes a handful of chopped-up bait in his left hand and thrusts it about 18 ins. under water. (He is facing the stern during the whole operation.) Then, reaching back with his right hand, he takes the kausele with the noose from tāfā, and immersing it in the water, thrusts down the tip of the kausele either to right or left of the slack of the hanging part of the noose. This slack then becomes the upper half of the noose while the running-line becomes the bottom half. The noose is thrust down between the handful of bait and the canoe, and is made to lie in a vertical plane parallel to the side. It is so manoeuvred that the after edge of the loop is about one span from the bait. The tautai then draws out the tip of the kausele, and casts it either over his right shoulder or under his left armpit to the tāfā. 16

- 60

The whole operation, from the time the pala is first seen takes but a few seconds. Immediately the kausele is withdrawn, the tautai calls to the bow paddlers “Ke mau atu tou vaka!” (“Hold your canoe off!”) These, with one quick stroke, pull the canoe forward, leaving the loop and bait astern. The line is allowed to run out sufficiently to prevent the noose from moving.

When the pala has devoured all the bait thrown out on the port quarter, the handful which was thrust down to starboard attracts his attention and he begins to move slowly towards it. While he is still some little distance (this varies for different tautai from a few inches to two feet) from the loop, the tautai, who has been leaning astern with outstretched right arm holding the line, makes a sudden overarm swing which causes him to fall back against tāfā The bow paddlers put pressure on their paddles to prevent the canoe from recoiling. The pala, startled by the movement of the noose, makes a lightning dive through it in the direction of the bait, but the noose closing on his body becomes fast (lave) in the depression in front of his stiff caudal fin.

Now there ensues a stern struggle. If the line is strong, the fish is given very little play. The tautai, who quickly recovers his balance, the tāfā and the tino loto, all haul on the line, and, as the pala, with a mighty splashing, is drawn up tail first by the tautai, tāfā seizes the braining-club (siki) and beats its head until it ceases to struggle. It is then thrown in the bottom of the canoe.

All apparatus is at once put in order again and an endeavour is made, by casting the bonito-head, as before, to entice another pala up to the canoe. If this fails, the trolling line is put out again and the canoe moves off to another place.

The average catch of pala in good weather is three or four for each canoe. I have seen as many as twelve brought in by one canoe.

All pala-fishing canoes return to the village in the late afternoon.

- 61

FLYING-FISH are caught both by dipnet (tae) and by a large folding-net (ikupenga). The use of the latter is becoming rare. It is said that in former days they were obtained sometimes by trolling with a leaf-lure, in which was a short piece of the midrib of a coconut-leaf pinnule (tua-niu), sprung into the shape of an arc, to the centre of which was attached the end of the trolling-line. The fish, taking the lure, released the ends of the tua niu which were held in position by being imbedded in the leaf, and these, straightening out transversely with regard to the trolling-line, became engaged in its mouth. 17

The fish are caught in large numbers in the short twilight that succeeds sunset. It appears that, during this time, they are unwilling either to fly out of the water or to swim more than a few feet below the surface, and it is in consequence of this peculiarity that they may be driven by canoes in any direction and herded in thick shoals into pockets on the outer edge of the reef. Here they are lifted in dipnets and flung into the canoes. It is during this sunset flying-fish hunting that the folding-net (ikupenga) is sometimes used.

Flying-fish are also caught in dipnets by torchlight (lama) before moonrise on nights when the moon is waning, or after moonset on nights when it is waxing.

It is usual, when the sea is moderately calm on the lee shore, for the majority of the village canoes to engage in the sunset herding of flying-fish. Some few minutes before sunset they are paddled through the breakers to the open sea, where they separate into several groups at intervals of a half-mile or so along the coast. Each group may have - 62 from two to ten or more canoes, all of which line up at intervals of about twenty fathoms, with their sterns a few fathoms from the edge of the reef and their bows pointing out to sea. The bow paddler in each canoe stands and searches the ocean to his immediate front for that moving shadow of a surface-disturbance which indicates a shoal of flying-fish swimming in to play at the reef-edge.

Immediately such a disturbance is sighted, two of the nearest canoes set out, and, one on each side of the shoal, taking care to keep slightly astern of the leading fish, they drive it in, the bow paddlers guiding it in the required direction by beating on the hulls of their canoes and splashing the water vigorously with their paddles. The other canoes of that group, or indeed any other canoes which may happen to be in the vicinity, now close in behind the shoal, and with great commotion assist in the drive. The fish progress in a very thick formation, blindly following the leaders, and at a rate of about two knots. I have seen shoals thus driven in from a distance of at least two hundred fathoms from the reef.

On approaching the reef, the fish bunch still closer together, and at the head of the shoal they appear to be a solid mass. They swim usually at a depth of about one foot, but ever and anon fins appear above the surface. No fish attempts to fly, nor will any dive unless the shoal is prematurely disturbed by over-eager hand-nets while it is still in deep water. When near the edge of the reef an attempt is made to head the shoal into a shallow V-shaped indentation (ava) from which it is difficult for them to escape. It is now the privilege of the bow-paddler in one of the two leading canoes, usually he who first discovered the shoal, to make the first dip. He thrusts the bag of his long-handled net into the shoal a few feet behind the leading fish. Quickly turning it mouth upwards, he draws it toward him and at the same time heaves it out of the water, usually full almost to the brim with struggling fish. All the nets within reach are by this time plunging in. Those which retrieve a considerable number of fish are swung round to be emptied carefully into the basket (tapola) amidships in each canoe. Those, however, which retrieve only one or two at a time, are flashed upwards with a lightning overhead swing which, stopped short when over the right shoulder, - 63 dashes the fish straight into the basket from a distance of at least ten feet. The accuracy displayed and sustained in this movement is most remarkable. It is but seldom a fish is misdirected so far even as to strike the side of the canoe, and the impact with which it falls is usually sufficient to stun it. The mêlée and the uproar are indescribable. Everyone is shouting directions to the netters, and the netters are shouting to the tautai 18 and the paddlers. On occasions a canoe will capsize, much to the amusement of everyone, including the capsized crew themselves. If the tide is high, many of the fish will have darted through the breakers on to the flat reef. These are pursued by canoes at great risk of swamping, the bow-paddlers standing all the while, preserving perfect equilibrium, not unlike slack-wire artists, with their long-handled dipnets for balancing-poles. The shoal has long since broken up into small groups of fish which dart about blindly, pursued by individual canoes.

At low tide, when the sea is at all rough, it is no uncommon sight to see one or more canoes perched on the very topmost crest of a rising comber, their bows jutting out over the sheer cliff of water and overhanging the bare jagged reef some ten feet below while the bow-paddlers stand and pluck flying-fish with their nets from the transparent wall of water. The tautai, with uncanny precision, avoids the crash by a few strong backwater strokes which send the canoe sliding to a hazardous safety down the smooth sloping back of the comber. Occasionally a canoe is too slow in sliding back, or is caught by the quick upheaval of the succeeding wave, and in spite of the mad efforts of the crew, with suddenly reversed positions, paddling furiously astern, is unable to mount to the summit before the crest curves over. On such occasions the canoe is dashed to pieces, but the occupants, diving deep into the base of the wave, always escape unhurt.

It is usual to work more than one shoal, and when the flying-fish are abundant as many as five or six shoals may be worked before it becomes too dark. The numbers caught on these sunset expeditions average about twenty fish per canoe, but it is by no means exceptional for individual canoes to capture a hundred or more.

- 64

Catching flying-fish in the folding-net (ikupenga) requires the concerted action of at least ten canoes, and it is therefore a pre-arranged communal expedition. The ikupenga itself is a net about 8 fathoms in length by 1½ or 2 fathoms in depth. Its upper edge is supported throughout its length by a boom of light wood jointed in the centre in such a manner as to allow it to be folded together lengthwise. The lower edge of the net is similarly attached to a folding-boom which is weighted sufficiently to make the net hang perpendicularly in the water.

The net, folded and rolled up, is carried out to sea on the outrigger-platform of the canoe of an experienced tautai.

When a shoal of flying-fish is sighted, several of the canoes set out to drive it in, guiding it in the manner already described. Meanwhile, two canoes remain with the net which is now cast overboard, unrolled, and unfolded into the shape of a V. Only one paddler remains in each of the two canoes; the remainder plunge overboard to attend to the net, those of one canoe holding in position the outer end on one side, while those of the second canoe attach themselves similarly to the opposite side. As the shoal is driven in, all the remaining canoes form up in a gradually contracting major arc of a circle abutting on the ends of the V-shaped net. All beat on the drum-like sides of their canoes with their paddles and shower spray gently on the straggling after-end of the shoal.

As soon as the leading fish of the shoal have headed into the net, the word is given by one of those in the water holding the net, and, with a great shout, all the crew but one of every canoe jump into the sea, and with threshing arms flail the water into a semicircle of white fury. The shoal of flying-fish darts forward into the apex of the net. The men at the extremities of the booms swim together, thus closing the folding sides. All others who are near dive down to the lower booms, draw them together and secure them by rolling them over once or twice on the doubled net.

Any fish which may have been startled into flight by the first wild plunge are quickly caught in out of the air by the lightning swings of the long-handled dipnets wielded by those who remained aboard the canoes.

- 65

The net, with all its fish, is lifted into the principal canoe, where it is emptied of its contents. It is then thrown back into the water and prepared to receive another shoal.

It is interesting to note that it is strictly tapu for anyone to lay a hand on the gunwale of the principal canoe, except, of course, the crew of that canoe themselves. 19

During the afternoons of those days on which the absence of moonlight in the evening makes torchlight fishing possible, the women and children collect the dry fallen coconut leaves and bind them together into torches. Each torch is about nine feet in length and is composed of three leaves held tightly together by the pinnules of the outer leaf knotted together at intervals of about one foot. The number of torches prepared for each canoe depends upon the time available for fishing before moonrise or before sunrise as the case may be. It seems usual to estimate about four torches per hour. After sunset the torches are borne down to the beach and deposited near the canoes.

The tautai as well as the bow-paddler of each canoe brings a dipnet. This is a shallow bag of netting having a mesh of about one inch, and supported around the edge by a stiff frame of ngie wood (Pemphis acidula). Two trimmed branches of ngie, having an average diameter of about ½ in., are lashed, one on either side, some inches in from the thinner end of a tapered handle of milo wood (Thespasia populnea?) about two fathoms in length by 1½ ins. diameter. To the end of this handle a cross-piece about six inches in length is attached. The ngie branches are lashed again to this cross-piece so that they spread out fan-wise beyond the extremity of the handle. Their fine ends are then drawn together and lashed to form a scarf-joint. The completed frame has the shape of a pointed oval, about 2½ ft. in length - 66 by 1½ ft. in breadth. The periphery of the bag-net is laced to this, and usually hangs so that it is shallow near the handle and deeper towards the outer extremity. The whole dipnet thus conforms closely, except in the disproportionate size of the handle, to the shape of a mandoline.

After the canoe has been carried down to the water and floated, the torches are placed on the outrigger-platform. The crew then gets aboard and the torch-bearer (he who occupies the seat immediately behind the bow-paddler), stands and takes all the torches, and placing them butts down in the hull, supports them vertically in front of him to protect them from being splashed by the breakers until the canoe has passed the surf-line on the outer edge of the reef. They are then replaced on the outrigger-platform.

The canoes draw up in line (tamanga) facing north, so as to sweep a front parallel with the reef on the lee side of the island from the village, which is near the southwestern point, to the other extremity of the lee reef. The canoes preserve an average interval in line of about three fathoms.

If more than ten canoes put out, it is usual to form two tamanga, one of which proceeds first, to be followed by the other at a distance of about ½ mile.

When all is ready, and darkness has fallen completely, one of the senior tautai gives the word to light up, and when the torch of his canoe is blazing, those of the canoes on either side are lighted from it and the light is passed on from one canoe to the other to the end of the line.

The scene is one of indescribable splendour, the village fires in the distance making dots of light in the palm jungle which lies in black silhouette against the skyline, the surface of the water studded with numerous little balls of phosphorescence which sparkle over the inky blackness of the depths, and the powerful glare of blazing torches overhead and around.

When all are ready, the line commences to move forward at a steady rate of about two knots.

In the bow of each canoe stands the bow-paddler with his dipnet held horizontally across his front so that the bag is to starboard—the clear side of the vessel. 20 That member - 67 of the crew immediately behind the netter in the bow stands and holds aloft the blazing torch. It is his duty to keep the torch well trimmed and burning brightly. This he does by reaching up, when necessary, and untying the bands which hold the leaves tightly together, and also by slapping vigorously with his hand on the lower part, thus causing the head to shower forth sparks and ashes. The other members of the crew and the tautai (captain and steersman) ply their paddles and keep a sharp lookout for flying-fish lying or swimming near the surface in the vicinity of the canoe. These appear light grey in colour in the glare of the torches, and are easily seen at distances up to about forty feet. Sometimes, when exhausted after a long flight, they lie motionless at the surface with wing-fins spread out as in flying. At other times they swim in shoals across the bow of the canoe within a few inches of the surface, usually proceeding seawards from the reef. When a flying-fish is seen, there is a general shout to the netter indicating the direction of the fish from the canoe, such as “Ama!” (port), or “Taumua!” (ahead). 21 Usually, however, the netter himself is the first to see the fish, and if it is out of range of his dipnet he calls to the paddlers to urge the canoe ahead, “Mumea! Mumea atu!” (A little further! A little further yet!) or shouts to the tautai to change his course slightly to port or starboard as required.

When the fish are running well the uproar is deafening. To the neophyte all seems chaos, but there is a definite thread of order and arrangement. Each canoe keeps its place in line. The outer ends of the line are a little ahead of the centre. Periodically, at the command of the senior tautai, a change takes place in the order of the canoes in the line, those which were on the inside taking position at the extremities, and vice versa. There is rarely a collision or a capsize, though it is not uncommon, when the sea is at all rough, to see several natives plunge overboard on the outrigger side of their vessel in order to seize and bring back to the surface an outrigger float which has sunk so far that the top-weight of the canoe, following it to that side, proves too great for its natural buoyancy.

- 68

Usually the netter scoops up one fish at a time and flings it into the basket in the same dexterous way that was described above under the heading of tae-isave-ao. In the thick of a shoal, however, a netter may capture as many as three fish, one after the other, before emptying his net. In order to do this, he twists the handle of the dipnet, giving the bag at the extremity a spinning motion while it is in the air so that any fish already in are kept by centrifugal force in the bottom of the bag, and are prevented from jumping out through the wide aperture while the netter is manoeuvring for position to scoop up another fish.

Shoals of large fish will on occasion follow the canoes in order to prey on the dazzled and exhausted flying-fish which escape the notice of the netters. At times, when these (especially if sharks) are ravenous enough to come up and compete with the canoes, whole shoals of flying-fish may be startled into sudden flight. Then it is not infrequent for the fishers to be struck with considerable force by fish flying blindly. Numbers fall into the canoes, while the netters with deft strokes of the dipnets, pluck them out of the air and dash them into the open baskets.

As stated previously, the tautai, as well as the bow-paddler, is equipped with a net. He, however, leaves his net lying on the lango kofe 22 of the outrigger-platform. Should it chance that his attention be directed to a fish which has escaped the netter in the bow and is approaching the stern, or to one which, for any reason, has come within easy reach of his position, he casts his steering-paddle into the canoe, and, rising simultaneously and seizing his dipnet, attempts to catch the fish. Whether successful or not, unless the fish are approaching the stern in numbers, he immediately returns his dipnet to the lango kofe and resumes paddling.

It may be of importance for comparative purposes to note some of the methods of handling the dipnet. Two points have already been mentioned, viz. the methods of casting the fish into the basket, and the spinning of the bag in order to prevent the fish from escaping. Another point worthy of notice is the method of handling the net in order to secure the fish in the bag. When the fish is lying dazed on the - 69 surface of the water, the mouth of the net is brought down flat on the water with a resounding slap, and in such a manner that the periphery of the bag surrounds the fish which, startled into flight, leaps up into the slack of the bag. Immediately after the slap, the mouth of the bag is twisted quickly upward and lifted clear of the water with the fish inside. This is the quickest method of taking the fish from the water, but can be used only with fish which are right on the surface. When the fish is swimming, the mouth of the bag is thrust under water a few inches in front of it, in such a manner that it will, unless it changes its course, swim straight into the opening. As soon as the fish enters the bag, the fisher, in one motion, twists the mouth upward and hauls straight in toward him. This is the quickest method of drawing the net out of the water after it has been immersed more than a few inches below the surface. An attempt to lift it straight out would put too much strain on the light shaft which forms the handle. It is unusual to attempt to capture fish swimming more than a few feet below the surface, but I have, on occasions, seen them brought up from a depth of a least eight feet. The net is wielded on both sides of the canoe, but it is good practice so to manoeuvre the vessel that the fish presents itself on the starboard bow.

Ceremonial expeditions and village competitions are sometimes arranged. On these occasions the return of the tamanga is awaited by a large portion of the population on the village malae. A fire is built in the light of which a count is made of the fish brought in by each canoe, and the total is announced by a herald in a voice loud enough to be heard throughout the village, “Te vaka o Penaia, laua, mo Luao, e lama foki; e mata-lau mata-iva-ngafulu ma fitu ko mate!” (Penaia's canoe, he and Luao, also fished by torchlight; one hundred and ninety-seven were caught!). The baskets are so counted five at a time, and, on the completion of a batch, the herald demands of the audience, who are expected to have memorised and computed the totals, “Taofi o koutou i vaka e lima?” (Your opinions concerning the total of the five canoes?). If several different totals are received, that given by the majority is accepted as correct. The result of the competition is announced by the herald and acclaimed by the adherents of the winning side with - 70 the loud shout of victory, “Uluf o-o-onu! Ui-e-e!” (Ulu fonu, lit. Turtle's head), while the young folk prance about on the malae, throwing handfuls of sand into the air and gibing at the losers. This is one of the most impressive old-time ceremonies still in existence.

The official counters toss the fish out of the baskets in pairs on to a common heap which sometimes contains, on completion, more than 3, 000 fish. Each pair is counted aloud, “E tasi-la, e lua-la, e tolu-la, e fa-la, e lima-la, e ono-la, e fitu-la, e valu-la, e iva-la, katoa-la. 23 The pair accompanying the word katoa is flung down at the feet of the counter, and, when the basket is empty, the number of fish lying at his feet indicates the number of tens counted out. To this total is added the number of units, if any, not sufficient to complete a final ten.

I have heard official counters, in all seriousness, but somewhat to the amusement of the onlookers, counting glibly aloud but using the names of a series of adjacent islands in place of the numerals. One common count utilizing this method commences with the two southernmost islands of the Gilbert Group, Tamana and Arorae, and proceeds southward through the eight islands of the Ellice Group.

On the completion of the announcements, the fish are divided out among the family groups in proportions which seem to take into account the social importance of the head of the family, the size of the family, and the value of the family's contribution to the communal effort.

- 71

IT was found inconvenient to make a study of the canoes of all islands of the Ellice Group, so, for the purposes of this paper, only those of one island, viz. Vaitupu, were studied in detail. Owing largely to the facilities of interisland communication of latter years, the differentiations in construction that may have developed on isolated islands are now to be found in the various styles in use among different families on any one island. The island of Nui, the people of which are descended from Gilbertese stock, is excepted from the above statement. Gilbertese innovations may, through communication with Nui, have been incorporated into the true Ellice Islands canoe, but it has been impossible to estimate the extent of Gilbertese influence, since no opportunity has occurred of studying the construction of the Nui Island canoes.

The large sailing-canoes of ancient times (lualua and foulua) are now no longer made. Tradition merely preserves incidental details of their construction such as that method of lacing the planks known as fou take. The canoes at present in use on Vaitupu are dugouts, and may be classified in three divisions. The local names will serve readily to distinguish the one from the other.

• (i) The Vaitupu type. According to local builders this is the traditional type of the Southern Islands.
• (ii) The Nanumea type. Named after the northernmost island of the group. This type which is now becoming popular in the southern islands is said to be the traditional type of the northern islands of the group.
• (iii) The paopao. Paopao, in the local dialect, signifies a small canoe of any type. Paopao, except in the few instances in which they are intended for use in the open sea, are roughly and somewhat carelessly constructed.

Trees of suitable proportions and possessing wood of sufficient durability out of which hulls may be fashioned are extremely scarce on all the lagoon islands. The only - 72 trees used in the Ellice Group are te fetau (Callophyllum inophyllum) and te puka (Hernandia peltata). Kanava (Cordia subcordata), which is used in the neighbouring Union Group, does not, in the Ellice Islands, attain suitable proportions for canoe-building. Puka, although not so durable, is greatly preferred to the brittle and heavy fetau, on account of the ease with which it may be worked, its lightness, and its ability to stand rough usage on the reef without cracking. Breadfruit logs have been used successfully, but for obvious reasons their use could not become general.

In ancient times on Vaitupu, it is said, the making of a new canoe was an event of some considerable importance. The skilled woodworker (tufunga) of the fortunate family possessing the suitable tree would call in the aid of the tufunga of allied families to the number of ten. On the day before that set for the felling of the tree, each tufunga would collect and sharpen as many shell and stone adzes (toki fasua and toki uli) as he could muster. On the following day, all the tufunga with other members of their families would proceed to the tree. The members of the family owning the tree would not be present, as it was their duty to provide food for the kau tufunga.

The tree was felled by making a series of parallel vertical incisions on opposite sides of the trunk and then chipping away with adzes the solid ridges between these. When the tree had fallen and the log had been trimmed, the leading tufunga would decide on which side the keel was to be. This side was turned under and the upper surface (te alo) was then trimmed down to what was subsequently to become the gunwale of the canoe (te ngutu o te vaka). When this had been accomplished, the kau tufunga had a rough estimate of the line on which they were to trim the outside of the body (te tua). The log was now turned over so that the keel was uppermost, and was raised a few inches from the ground by two transverse supports called lango. The kau tufunga took up their positions, one at each end and one or more on each side. The leading tufunga indicated port (ama or outrigger side) and starboard (katea). This was necessary as, although the central part of the body and the rounded keel of that part were trimmed straight with the log, the sharp keel line (kalisi) commencing about a - 73 fathom from the bow and terminating at the bow was set parallel with the medial line, but slightly towards the outrigger side, while the other kalisi, about three-quarters of a fathom in length, and terminating in the stern, was set similarly to the opposite or starboard side of the central line, Fig. 36. When a certain amount of rough trimming had been accomplished, the log was dragged to a convenient place in the village.

Early in the morning of the following day the kau tufunga again commenced work. The exterior was trimmed to its final shape. Here it might be noted that the ideal shape of the Ellice Islands canoe is that of the body of a whale. Some tufunga occasionally attempt to fashion a hull after the shape of the bonito (atu) or of the pāla, but the bodies of almost all canoes bear a strong resemblance to that of the whale (tafola). The shell and stone adzes rapidly became blunt, and each tufunga fashioning the body was assisted by another who was continually at work sharpening the edges of one adze after another as they were passed to him.

It was naturally to the advantage of the family providing the food to hasten the completion of the work, and in this connection it is interesting to note a traditional saying still quoted sententiously to the malingering tufunga.

When the outside of the hull had been trimmed to the desired shape, the log was turned over and the tufunga commenced to hollow out the interior (te liu). Extreme care was necessary for this part of the work, as a single false stroke with the adze might ruin the canoe. Large heavy adzes were used to hollow out roughly the central part of the log, leaving the sides about two inches thick. Small light adzes with a curved blade (atu pa—Northern Ellice dialect fakatipa) so arranged as to be reversible laterally for trimming either side or the bottom were then employed.

Some ceremonial attached to the conduct of the members of the kau tufunga during the course of the work. They lived together, and as stated, were fed by the family whose - 74 canoe they were building. Stipulated quantities of food had to be brought to them at certain times. A quantity of food would be brought as early as possible and offered to the leading tufunga who would pray (?) over it. This daily ceremony was referred to as lotu a toki (religious ceremony of the adzes). The tufunga did not Partake of this food, but ordered it away for distribution to their families. They then proceeded with the work. At about 9 a.m. (fangainga o lupe—lit. feeding of pigeons) the builders would make their breakfast of a second quantity of food provided. The work would then proceed. During the day two more quantities of food would be brought, one for a mid-day meal (te kainaki tutonu) and the other to be eaten on the completion of the day's work (kai o moe). Immediately after this the kau tufunga proceeded together to bathe.

Inquiries as to the religious aspect of the ceremony have failed to elicit any information, probably owing to the strict tapu imposed by the native Samoan missionaries who christianized the group, on all reference to the customs of ancient times.

In modern times, the procedure is in general the same. Two tufunga, however (using modern tools), are usually considered sufficient. The adzes used are made from planeirons bound to the old-style helve. The reversible adze mentioned above has the traditional helve, while the blade is made from that of a gouge-chisel. In trimming the interior sides and bottom (te liu), the proper thickness is judged by sound. The finger-nail is flicked against the place to be tested and the resulting sound compared with that obtained by flicking another similar place already finished satisfactorily. Sometimes one hand is held on the exterior face to feel the resonance when the interior face is flicked. The upper edges of the sides are made about 1½ ins. thick while the sides themselves are chipped down to little more than ½ in. The bottom again thickens out to 2 ins. or more at the keel-line.

It is rarely a log is obtained of such dimensions and so free from faults that the whole body of the canoe can be fashioned in one piece. The most common fault is a cleft or depression in the surface, which leaves a corresponding depression in the side of the canoe. When the depression occurs near the upper edge, the practice is to cut away the - 75 side at that place. The upper edge is then raised to the required height by superimposing and lacing down a topstrake of the same wood. These raised sides are called oa. The sides of a canoe made from a small log are always raised to the required height by means of oa. The fitting of the oa to the body is an operation requiring nice judgment and skilful workmanship.

The method of trimming any two surfaces to make an exact fit merits notice. The fitting surfaces are trimmed with the adze to the approximate shape. One surface is then smeared with a black mixture of powdered charcoal and water. This surface is then applied to the other, and, on their being separated, the projecting irregularities of the unsmeared surface are indicated by black marks. These black marks are cut away with the adze and the two surfaces are again placed together. This operation continues until one surface makes contact with the other at every point of its length.

On the Nanumea type of canoe, before oa are fitted and laced down the limits of the bow and stern covers (puke-mua and puke-tua) are marked. The puke-mua is usually given a length of one fathom (ngafa tasi) from the bow, while the puke-tua has a length of three-quarters of a fathom from the stern (fatuli—i.e. the distance from the finger tips of one extended arm to the flexed elbow joint of the other extended on the opposite side of the body). Further consideration of the two puke is deferred in order that they may be described in their proper place in the order of the assembly of the various parts of the canoe.

The oa, then, on the Nanumea type of canoe do not project, fore and aft, beyond points marked for the inner edges of the puke (ngutu-puke), but on the Vaitupu type, they may extend to within a few inches of the bow or stern.

The height of oa is determined approximately by the height of the knee-joint when the foot is placed on the bottom inside the body. Variations in this measurement, when intentional, are for the purpose of fitting the subsequent height of the seats to the stature of the persons who will normally form the crew. The natives hold the theory that endurance in paddling long distances after bonito is determined largely by a comfortable posture with legs neither too straight nor too cramped. A further consideration which - 76 tends to limit the height of the side is the fact that a canoe with high sides requires long-handled paddles, which are more difficult to wield than short-handled ones. In fitting the oa to the sides, both are given a very slight inclination towards the outrigger side in respect to the plane of the medial, longitudinal, vertical section of the body. This inclination, which is so slight as to be scarcely noticeable, is said to enable the joints of the oa with their respective sides better to withstand the strain of the outrigger-booms. A method of ensuring rigidity in these important seams is so to shape the fitting surface of the oa plank that when placed lightly in position it rests on its ends while the central postion shows a small space between the two surfaces. When the oa is laced down to the side there is, then, a permanent tension which helps in resisting the tendency to lateral play under the strain applied by the leverage of the outrigger-booms. Tradition states that in ancient times a strip of pandanus leaf was placed in the seam throughout its length, not so much to prevent leaking as to keep the seam itself dry and so prevent rot. This strip of leaf was replaced from time to time. Modern canoes rely on a coating of coal-tar on the outside of the joint and joint-lashing to perform the same service. In spite of this many modern canoes quickly become unserviceable through softening of the wood and consequent slackness in this region.

The oa is held in place before lashing by means of three string-and-wood clamps known as fakafiti, one at each end of the oa and one in the middle. A fakafiti consists of a loop of cord threaded through two lashing-holes in the side of the canoe. The inner and outer sides of the loop are drawn up and clamped over the top of the oa by means of two sticks each of which holds the other from slipping out of place. See Fig. 37.

The oa is lashed to the side by means of three-ply coconut sennit (tuli kafa) passed three (tua-tolu) or four (tua-fa) times through pairs of holes spaced a short thumb-to-index-finger span (anga foliki) between pairs. Usually the hole in the oa, in each pair, is situated vertically above the hole in the side and each is spaced about .7 in. from the seam. The sennit is threaded though the holes by means of a needle made from the stem of a coconut-leaf pinnule. Each turn is drawn tightly. The inner end is not knotted - 77 but is caught fast under the second turn. The outer end is drawn several times around the binding-turns in the region of the depression on the outer side of the seam. It is then knotted and neatly cut off. Finally the holes are plugged tightly with wooden pegs (pono) which hold the lashing securely and also prevent leakage.

Sometimes it is necessary, before fitting the main oa, to build up a piece of the side where it has been cut away on account of a deep depression or other fault in the material of the original log. This inserted piece is known as te fono. The word kautaka is applied to the projecting bulge on the upper edge of the body which does not lie in the same line as the outer surface of the oa (See Figs. 38 and 38a). The obtuse angle formed by the junction of a bulging side and the oa-plank, when it is so pronounced as to cause the binding-cords to stand away from the wood, is sometimes packed with a long sliver of puka wood (Fig. 39). The purpose of this, one tufunga has informed me, is to give the canoe a more finished appearance, but I am inclined to think, having regard to the meagre development of decorative art in the Ellice Group, that the taka-oa, as this sliver is called, was originally introduced to pack the lashing and to prevent the knotted outer end of the binding cord from working loose.

Since the order of assembly and the system of determining measurements are, in general, the same for both main types of canoe, I propose, in this account of the making and assembling of the various parts, to confine the remainder of the description to the Vaitupu, or southern islands type and, later, to consider the essential points of difference in the Nanumea, or northern islands type.

At this point in the construction, the canoe has the appearance shown in the outline sketches in Figs. 41 and 42.

It may be well to remark here the shape of the sides of the body as shown in the cross-sectional sketch in Fig. 40. It will be noticed that the lateral bulge (inaki) on the outrigger side is in a lower horizontal plane than the bulge (inaki) on the starboard side. The explanation of this is that the former is usually the natural bulge of the original log while the latter is purposely made as high as possible in order to resist any tendency to capsize to starboard (mafuli), an accident which frequently happens at exciting moments - 78 when all the crew inadvertently lean over to starboard to watch or take part in some fishing operation.

With regard to the shape of the body, it is necessary to record another point which easily escapes casual observation. A horizontal section at any level within several inches of the waterline of a well-built canoe would show a slightly greater curvature on the outrigger side than on the starboard side. Reference to Fig. 36 will show more clearly what is meant.

I think that this pronounced curvature of the longitudinal are of the port side and the flattening of that of the starboard side really explains the difference in the levels of the cross-sectional bulges or inaki shown in Fig. 40, although tufunga on the island of Vaitupu have found the convenient explanation given above.

It is interesting to note that the fast, carvel-built sailing canoes of the Gilbert Islands have the curvature on the outrigger side so pronounced as to give the body an appearance of asymmetry.

The reason for this is obvious. A small piece of wood cut in this shape, but having no outrigger attached, will, when pushed through water, persistently steer towards the straighter side. The body of the canoe therefore has a tendency always to steer away from the outrigger, which tendency in a well-built canoe is exactly counterbalanced by the drag of the outrigger itself.

To proceed with the assembly, the tufunga, having fitted and laced down the oa, next prepares the booms (kiato) and float (ama) for the outrigger. All present-day canoes have the boom, and what might be called the boom-leg (tapuvae) in one piece. This is obtained from a bough and branch of pua (Guetarda speciosa) or tausunu (Taunefortia argentea). The bough must be fairly straight for about a fathom of its length, with a diameter of about 3 ins., and with a substantial straight branch at the outer end. The angle of the branch with the bough may vary for different canoes between approximately 120 and 160 degrees, but the branchangles of all boughs chosen to make the booms of one canoe are the same. The method adopted by one tufunga on Vaitupu to ensure the approximate equality of these angles is to cut first a small bough of ngasu (Scaevola Koenigii) having a branch at the angle he requires. This he uses to - 79 gauge the branch-angles of such boughs as he may discover in the bush. The number of booms required for a canoe varies from 1 to 5 according to size. Both extremes are rare. The usual numbers employed on ocean-going canoes are 3 and 4, while lagoon paopao are made commonly with 2 or 3. A canoe having 3 booms requires, normally, 4 paddlers (ta-fa), while one with 4 booms requires 5 paddlers (ta-lima).

The outrigger-float is made almost invariably of puka (Hernandia peltata). It may be cut from a bough of the tree from which the body of the canoe is fashioned, but it is sometimes necessary to cut another small puka tree.

The float is trimmed to a maximum diameter of from 5 to 9 ins. according to the size of the canoe. It is usually given such a length that the forward end will, when in place, be laterally opposite the feet of the bow-paddler (tino i mua), and the after end opposite those of the fishing captain (tautai) who occupies the seat in the stern. The general shape of the float can best be illustrated by diagram (Fig. 43a, b, c and d).

When the float has been trimmed roughly to shape, it is set in position on the port side parallel with the body and distant from it about three-quarters of a fathom (fatuli—see above). The forward end is then set in towards the body a short distance that varies with different makers from a few fingers' breadths to a span. Thus in its final position the lateral space between the keel and the outrigger at its after end exceeds its distance from the keel at the forward end. This is a most important point in the construction, as it determines the ease with which the canoe may be turned to starboard when at sea. Reference to the tables of measurement given below will show that of 13 canoes chosen at random from among the canoes at the main village on Vaitupu, all but one (an unimportant lagoon paopao) have the forward end of the float nearer to the keel than the after end. I have personally experienced that a canoe with parallel outrigger float is extraordinarily difficult to steer to starboard, although it is quite as easy to steer to port as one with a properly set float. The parallel float, theoretically, offers less resistance to forward motion, but one of the chief requirements in a fishing-canoe is ease of handling, especially when one of the almost daily uses to which it is put is - 80 the hunting of flying-fish at sunset in a precarious position within a few feet of the edge of the reef.

When the float has been set in place, the leg-end (tapu-vae) of the forward boom (kiato) is cut so that it is about equal in length to the interior depth of the body at the place the boom will subsequently occupy. The body of the canoe is set exactly upright on its keel. Then, from the bow, the positions of various objects of the superstructure are marked.

This is the system used by one tufunga in marking the various positions in, for example, a 4-paddle canoe (vakata-fa) i.e. a canoe having three booms.

One fathom from the bow a mark is made for the inner end of the bow-cover (ngutu o te puke-mua). Proceeding aft from this mark, another mark is made at a distance of two thumb-joints for the manu mua (a transverse stick used as a grip in lifting the canoe).

Three spans aft of the mark for the manu mua is the position of the forward boom (kiato mua) at present under consideration.

Three-quarters of a fathom (fatuli) from the stern is marked the inner end of the stern-cover (ngutu o te puketua). One hand's breadth plus extended thumb (tapalima) forward of this the position of the fishing-captain's seat (nofoanga o tautai) is marked. Three spans forward of this gives the position of the manu tua (a transverse stick used in lifting the canoe), while one hand's breadth plus extended thumb forward of the manu tua is the position of the after boom (kiato muli or kiato ta-fa).

Slightly forward of the central spot between forward and after-booms is marked the position of the central boom. The reason for placing the central boom forward of the central spot is to allow plenty of room to the paddler who occupies the seat on the after-boom whose duty it is to operate the bailer (te asu).

To return to the fitting of the forward boom, the supports (lango) are removed from beneath the body of the canoe so that both body and float are resting on the ground, and the body is given a scarcely noticeable inclination from the vertical towards the outrigger. The purpose of this inclination is to set the outrigger booms at a right angle to the vertical plane of the body so that they will be parallel - 81 to the plane of the waterline when the crew is aboard. A well-built canoe, when floating empty, has a cant towards the outrigger. When the crew go agoard, the subsidence of the hull caused by their weight brings the booms into a horizontal position. The amount of cant allowed in thus fixing the outrigger is determined by the length of the tapuvae and varies considerably. It can be judged by the tufunga only by estimating the probable displacement. The horizontal part of the boom is then placed in the position already marked for it. To make it rest on both sides of the canoe, it is usually necessary to pare the under surface of the boom so that it tapers towards the starboard end.

The after-boom (kiato muli) is next fitted. Its leg (tapuvae) is cut to fit exactly on the upper surface of the after end of the float when the boom is resting evenly on both gunwales of the canoe. The inner boom is then similarly fitted. The upper surfaces of all booms are kept as nearly as possible in one plane so that the superstructure (kaupalepale) and carrying-piece (saunga) will lie evenly along them and so permit of firm binding in this region, a point of great importance to the strength of the whole outrigger.

The booms are now lashed to both gunwales. Usually they are kept square with the medial line, but on some canoes they are set forward at the outer ends so that the angle abaft the boom with the medial line of the body is slightly obtuse. This forward set is scarcely noticeable, but in the cases in which it occurs it is not accidental. One tufunga has stated that it assists the canoe-captain (tautai) to keep his course against the buffetting of a head sea. Immediately beneath the boom, in the side of the canoe at a distance of about two finger-joints (e lua fatinga a maikao) is drilled a hole of sufficient diameter to contain six thicknesses of lashing cord. If the tufunga's boring instrument (now-a-days a gimlet) is of small diameter, two holes may be bored at the same distance and spaced laterally by about the thickness of the boom. The lashing is done with threeplait coconut sennit. One end is held on the upper surface of the boom and the free end is fastened to the midrib of a coconut-leaf pinnule and threaded from the inside through the hole in the side of the canoe. It is then drawn tightly up over the boom, catching and holding its own - 82 inner end as it passes down on the other side of the boom, to be threaded again through the hole. The second turn is brought up on the opposite side of the boom from the first turn, and so with alternate turns, so that on the completion of the lashing, the crossed cords on the upper surface of the boom present the appearance of a neat, woven figure (i.e. e sumu or e manu tua o to kiato). Six turns (tua ono) is the usual number. The lashing is finished by taking two transverse turns round the joint so as to enclose all the vertical lashing-turns, and then several turns around the boom itself, the last of which pass over and secure the free end of the cord. In those cases referred to above in which two holes are bored in the side of the canoe for each boom-lashing, the turns seen from the side have a forked appearance (See Fig. 44b) and this style of lashing is consequently known as manga-lua (forked).

The next operation is the lashing of the float to the ends of the boom-legs. There are two methods by which this may be accomplished. The first and usual method is by means of pegs (suki, tuki or tona) driven into holes around the periphery of the lower end of the boom-leg which is then lashed to them with three-plait sennit. The second method is by drilling a downward slanting hole from each side of the float in such a way that the two holes meet immediately under the boom-leg, which is then lashed by many turns of cord passed alternately round the leg and through the drilled passage in the float. This method is seldom used now-a-days.

For the pegging method, one, two or three pegs may be used for each boom-leg. The traditional number for the northern islands style of canoe was three, and that for the southern islands two. I have seen one example of a strong ocean-going canoe utilizing only one peg for each boom-leg (Fig. 46). Typical arrangements of pegs are shown in Fig. 45a, b and c. The heads of the pegs are made to slope from the perpendicular outwards in respect to the boom-leg; this prevents the lashing from working upwards. It has also the effect, I presume, although the tufunga do not say so, of wedging the pegs and so preventing their becoming loose in their sockets. The lashing of the boom-leg to the pegs is first by the figure-of-eight method, each peg in turn with - 83 the boom-leg. The number of turns varies with the thickness and strength of the lashing-cord, which is of three-plait sennit. Usually the lashing is completed by several turns taken around all the pegs. The free end is bound several times around the boom-leg and secured in the usual manner.

In the second method, employing direct lashing (Fig. 47), the lashing-cord is bound several times round the boom-leg, then taken through the drilled passage in the float, again several times round the boom-leg and back through the drilled passage in the opposite direction. This is varied with a criss-cross lashing designed to cover the whole of the lower end of the boom-leg and hold it firmly on the flat upper surface of the float. Thus it tends to prevent that lateral motion of the float with respect to the boom-legs which in most cases loosens the lashing and necessitates re-lashing every three or four months. Some canoes employ both methods, one or more boom-legs being lashed directly by means of holes drilled in the float and the remainder being lashed to pegs.

The small lifting-grips (manu), mentioned above and shown in Fig. 48, are often set in place before the booms. They are lashed to the gunwales in a manner identical with that of the lashing of the booms.

These lifting-grips are usually made of small branches of the milo tree (Thespasia populnea).

The tufunga next turns his attention to the preparation and attachment of the outrigger lifting-piece (saunga) (See Fig. 48). This is made of milo wood or pua (Guetarda speciosa). It should be strong enough to bear the weight of the outrigger and long enough to afford a comfortable grip for the left hands of the carriers while their right hands are gripping the lifting-pieces attached to the body of the canoe. It is lashed securely to the outrigger-booms at a distance from the port side of the body that varies with individual requirements from about 14 ins. to 21 ins. The lashing to each boom is composed usually of six turns taken figure-of-eight fashion round both kiato and saunga.

After the saunga is in place a number of small straight branches of the pua tree, having an average diameter of less than one inch are bound to the booms, on the farther side of the saunga from the body, at intervals that may vary from less than one inch to several inches. These small - 84 sticks form a shelf, or platform, on which fishing lines and other light gear may conveniently be carried. They are named kaupalepale.

Sometimes it is necessary to raise the level of the upper surface of one or more booms where they curve away into their respective boom-legs in order to present a level surface for the attachment of the kaupalepale. This is done by the insertion of a wedge-shaped piece of pua wood known as to lango (See Fig. 49).

The outrigger is completed by the attachment of two vertical forks of tiale (Gardenia taitensis D.C.), ngie (Pemphis acidula) or milo (Thespasia populnea). These are best illustrated by diagram (Fig. 46). One is placed at each end of the saunga at the junction of the latter with the outrigger boom, to both of which pieces it is securely lashed. They are known as lango kofe, and their purpose is to hold the long rod (kofe) used for bonito-fishing, or the long-handled net (tae) used for catching flying-fish. Their extreme height does not usually exceed 1 ft., while the space between the extremities of the forks, which always have a wide angle, is about 9 ins.

The lifting-grip for the right hand of the person carrying the after end of the canoe is shown in Fig. 48, where it is marked (a). It will be noticed that the manu to which it is attached has been allowed to project a few inches beyond the port side of the body. The saunga (a) is lashed to the projecting end of this manu, and the distance from the side of the canoe is just sufficient to give easy clearance to the fingers using the grip. The forward end of the saunga is made to fit flush with the after boom in order to obviate all chance of interference with the free operation of fishing gear.

The lashing employed on this saunga is worthy of notice in that it was held, in ancient times, to have some considerable effect on the luck of the fishing. The actual lashing of the saunga to the manu was known as the li of the canoe. It was believed, and there is still some remnant of the superstition, that on the care with which this lashing was made depended the canoe's ability to attract fish. One old tufunga informed me that if the li were carelessly fashioned, such fish as were attracted to the canoe by the bait thrown out would devour the bait and swim away; but - 85 on the other hand, if the li lashing were made with due care, the fish having devoured the bait, would wait angrily for a chance to seize more and so were easily caught. The lashing itself, now-a-days, seems to differ in no wise from lashings already described except that some tufunga invariably use only one unbroken length of tuli kafa sennit to accomplish the three lashings on the after manu (i.e. the lashing to each side of the canoe and that to the saunga).

The bow and stern-covers (puke mua and puke tua) are now made. They are cut each from one solid piece of the same wood as that used for the body of the canoe. Different styles, and the measurements of the main parts, are shown in the diagrams, Figs. 50 to 53. Both puke mua and puke tua are attached to the body by tying-strings so that they can be taken off easily when ashore, in order that the canoe may be dried out thoroughly to prevent rot.

The main part of the puke mua is the wave-guard (pale ngalu). This was not found on Vaitupu canoes in former days, but has been borrowed, according to tufunga, from Nukulaelae, the southernmost inhabited island.

The raised portions of the puke tua named tokoulu, sake and tukungane are intended, the first for holding the bonito-fishing rod while the canoe is under way (See Fig. 53), and the others for holding prepared bait for line-fishing, etc.

Important differences in the bow and stern-covers of the Northern Islands' or Nanumea type of canoe will be discussed later.

The pointed knobs shown on some of the stern-covers are said to be shaped like a maiden's breasts, and are used in fishing for pāla as convenient projections on which to hang the line attached to the bonito-head which is thrown out to attract the fish.

The last operation is the placing of the canoe-captain's seat (te nofonga o tautai). This is sometimes “let in” (faka-to ifo) so that the upper surface is level with the gunwale. It is lashed to the canoe by one lashing at each end. These lashings are set rather towards the after-edge of the seat in order better to take the strain of the bonito-rod when this is set in the mata pili (See Fig. 48) . The lashing-cord is laced several times through two holes, one bored in the end of the seat and one in the side of the canoe.

- 86

To the after-edge of the tautai's seat is fastened the mata pili, a stout sennit ring made by winding about four turns of three-plait sennit (tuli kafa) around two or three fingers and then stiffening the ring by a single or double layer of seizing turns. This grummet is lashed to the tautai's seat by means of one or two holes drilled in the after-edge of the latter through which are passed several turns of lashing-cord. These lashing-turns are finally wrapped by several turns of the free end of the same cord. The mata pili is then held in position as regards its after-edge by means of guide-strings laced twice through small holes in the upper edge of the side of the body near the forward edge of the stern-cover (See Fig. 48).

The butt end of a bonito-rod is placed in the mata pili and rested against the highest projection (tokoulu) on the stern-cover, the slope is noted and a groove cut of dimensions sufficient to hold the bonito-rod in place.

A loose detachable seat made of puka, about .7 in. in thickness and from 4 to 6 ins. wide is sometimes placed on the forward side of each boom where it rests on the gunwales. It is seldom tied down.

This type of canoe is commonly found in the northern islands of the group where, owing to a dearth of puka trees (Hernandia peltata) it is usually made of the heavy, brittle fetau (Callophyllum inophyllum). As previously stated, it is now becoming popular in the southern islands where it is made usually of puka. On the island of Vaitupu, this type has so gained in popularity as to occur now-a-days in about equal proportions with the Vaitupu type. Figs. 58 and 59 show an ideal canoe of each type so arranged as to show clearly the principal points of difference in construction.

It will be noticed that all the main differences are in the superstructure. These may be enumerated as follows with initial reference to the Nanumea type.

• (1) The bow and stern-covers (puke mua and puke tua) have deep sides, and are laced permanently to the body.
• (2) The stern-cover (puke tua) has no tokoulu, or built-up rest for carrying the bonito fishing-rod at a trolling-angle while hunting for a shoal of bonito. The place of this is taken by a cross-piece (lango kofe) set immediately - 87 in front of the inner edge of the stern-cover. This crosspiece has a groove in the centre in which the rod lies during this fishing operation.
• (3) Extending from the raised inner edges of the bowcover to those of the stern-cover a long straight branch of fetau (Callophyllum inophyllum), trimmed to an even diameter of about two inches, is superimposed on each side. These are known as kaufuatanga. They cover the tautai's seat as well as all the outrigger-booms at the attachment of these pieces to the sides of the canoe. Each kaufuatanga is lashed at both ends as well as at the points of contact with the outrigger-booms, at which points the lashing is usually in one piece with that of the booms. The details of the lashing, varying as they do with different tufunga, merit no special notice. All lashings in this, as in the Vaitupu type, are usually made with the strong three-plait sennit (tuli kafa). The less durable twoply twist sennit (kolokolo) is sometimes seen in canoelashings but usually in unimportant places or in cases of emergency.
  It should here be noted that the lango kofe mentioned above is lashed on top of the kaufuatanga.
• (4) The upper surfaces of both bow and stern are trimmed to the general fish or whale-shape of the body. The stern is often forked like a caudal fin.
• (5) There are usually no special lifting grips (manu) on the hull, the place of these being taken by the kaufuatanga. When a manu tua is provided, it is often set obliquely as in Fig. 59.
• (6) It is traditional on Vaitupu, though not common practice in building, that the outrigger of the Nanumea type has three pegs to each boom-leg while that of the Vaitupu type has two (See Fig. 45 (c) for arrangement of pegs).

Here occurs a point of some interest. The kaupalepale, or platform on the outrigger, although it is identical on Nanumea canoes made on Vaitupu, with that described above for the Vaitupu type, is situated, on canoes actually built on some of the northern islands, between the outrigger lifting-grip (saunga) and the kaufuatanga on the port side of the canoe. This makes it impossible for the paddlers occupying the seats on the after and central booms to use their paddles on the outrigger side, and consequently renders them more liable to fatigue in paddling long distances.

- 88

Reference to the tables of actual measurements given below will show that this type of canoe is much smaller than that used for the open sea. It occurs in both Vaitupu and Nanumea types, although the latter is rare, since no bow and stern-covers are necessary on the calm waters of the lagoon. The photograph (Fig. 75) shows a typical paopao.

Paddles—Two distinct types of paddle occur on Vaitupu. A specimen of each type is shown in Figs. 54 and 55; (a) is the common type of paddle, while (b) is a type frequently used by steersmen (tautai). The chief woods used for paddles are pua (Guetarda speciosa), puka (Hernandia peltata), kanava (Cordia subcordata), tausunu (Taunefortia argentea), milo (Thespasia populnea), fou (Hibiscus) and fetau (Callophyllum inophyllum).

The Bailer—This is made of any of the woods enumerated above for paddles. All bailers are of the shovel shape illustrated in Fig. 56. The bailer in each canoe is operated by the paddler who occupies the seat on the after-boom. As mentioned above, the space between the after-boom and that immediately in front of it is always greater than the space or spaces between the other booms, in order that the member operating the bailer should have clearance for his arm as it sweeps forward and upward shovelling the water up and over the side.

In this connection, it is interesting to note that the phrase meaning “to bail out” is the same as that meaning “to hollow out the interior” in building the canoe, viz. o ta te liu.

Te siki—Each fishing-canoe carries a club (te siki). This is usually a branch of any heavy wood roughly trimmed and having a length of about 3 ft. and a thickness of some 2½ ins. It is used for battering the skull of a large fish before it is hauled into the canoe.

- 89

Principal measurements of five canoes of the Vaitupu type, five of the Nanumea type and three lagoon paopao, actually in use on Vaitupu in 1928.

Measurements in feet and inches.

No. of Canoe		1		2		3		4		5
	ft.	ins.	ft.	ins.	ft.	ins.	ft.	ins.	ft.	ins.
Length in feet	29	8	28	7.9	29	0.9	25	1.2	19	3.3
Maximum interior width of gunwale		14.9		13.3		14.1		13.5		13.2
Maximum interior width of body		18.2		16.9		20.1		15.5		14.2
Width of body at inner end of bow-cover.		13.4		13.5		16.1		13.2		10.9
Width of body at inner end of stern-cover		11.6		13.5		13.6		11.3		8.9
No. of kiato		4		4		4		4		3
Dist. of forward boom from bow	7	6	7	7.5	7	7.1	7	8.9	6	1.2
Distance of after boom from stern	9	3	9	4.4	8	6.9	8	2.9	6	10.9
Average space between booms	4		3	7	4	05	3	10	2	11.2
Angle of boom with tapuvae, in degrees		140		145		125		120		140
Length of central boom from port side of body to angle with tapuvae		22.5		26.6		38.5		36.1		31.7
Length of tapuvae.		25.5		24.7		21.7		18.2		19.7
Length of float	15	2	14	0.5	16		11	8.2	9	1.5
Greatest diameter of float		8.3		8.5		7		6.7		6.7
Lateral stance of float from keel—
(a) forward	4	6.9	4	6.3	4	8.2	4	2.7	4	3.6
(b) Aft	4	9.9	4	9	4	11.6	4	5	4	6.2
Space between saunga and port side of body		14		18.7		21.7		21.5		17.8
Bow to inner end of bow-cover	4	1	4	11.1	5	3.7	5	6.3	3	5.5
Stern to inner end of stern-cover	4	3	4	11.2	4	2.5	4	4.2	3	4.5
Depth of body—
(a) Central		19.8		20.6		20.9		18.2		15.1
(b) At inner end of bow-cover		17.5		17.5		17.8		15.2		14.5
(c) At inner end of stern-cover		14.7		16.1		15.4		12.6		12.8

- 90 Measurements in feet and inches.

No. of Canoe		1		2		3		4		5
	ft.	ins.	ft.	ins.	ft.	ins.	ft.	ins.	ft.	ins.
Length in feet	29	7	26	0.2	25	11.1	28	3.5	13	9.3
Maximum interior width of gunwale		14.2		12.7		12.5		13		11
Maximum interior width of body		18.2		14.8		16.2		15		12.1
Width of body at inner end of bow-cover.		12.2		12		11.1		11.6		9.6
Width of body at inner end of stern-cover		9.4		9.1		9.3		10		6.9
No. of kiato		4		3		3		4		2
Dist. of forward boom from bow	7	11	8	0.2	8	5.7	8	5.5	4	5
Distance of after boom from stern	9	1.5	9	10.5	9	0.2	9		4	4.7
Average space between booms	4		3	10	3	11.5	3	4	4	8.8
Angle of boom with tapuvae, in degrees		140		145		145		130		155
Length of central boom from port side of body to angle with tapuvae		27		26.3		30		29.7		25
Length of tapuvae.		20.9		25		23.2		23		20.5
Length of float	16	0.5	11	7.7	11		12	11.2	7	4.5
Greatest diameter of float		7.1		7.6		7.5		7.5		6.3
Lateral stance of float from keel—
(a) Forward	4	4.2	4	5.7	4	6	4	3.7	3	9.2
(b) Aft	4	9.9	5	1.6	4	10	4	9.3	4	1
Space between saunga and port side of body		17.7		17		20.2		19.5		17
Bow to inner end of bow-cover	5	1	4	3	4	5.3	4	7.4	3
Stern to inner end of stern-cover	3	1	2	9	3	8.8	3	3.2	2	2.9
Depth of body—
(a) Central		22.4		18.6		17.8		19		14.2
(b) At inner end of bow-cover		17.5		17		15.5		16.5		14
(c) At inner end of stern-cover		13		12.4		14		12.7		11

- 91 Measurements in feet and inches.

No. of Canoe		1		2		3
	ft.	ins.	ft.	ins.	ft.	ins.
Length in feet	15	7	17	8	14	5.2
Maximum interior width of gunwale		12		12.5		10.2
Maximum interior width of body		13.7		14.2		11.9
Width of body at inner end of bow-cover.		10		11.4		9.2
Width of body at inner end of stern-cover		9.2		10		6.4
No. of kiato		9.2		10		6.4
Distance of forward boom from bow	4	3.5	5	4	4	7.6
Distance of after boom from stern	5	5	6	0.2	4	5
Average space between booms	2	8	2	11	2	5.8
Angle of boom with tapuvae, in degrees		135		155		150
Length of central boom from port side of body to angle with tapuvae	27.2		25.7		27.9
Length of tapuvae.		20		20.3		14.2
Length of float	7	2	8	4.3	7	2.1
Greatest diameter of float		5.5		6.3		5.9
Lateral stance of float from keel—
(a) Forward	4	0.5	4	2.4	3	8
(b) Aft	4	1	4	2.3	3	9.7
Space between saunga and port side of body		15.4	—			16.3
Bow to inner end of bow-cover	—		—		3	1.3
Stern to inner end of stern-cover	—		—		1	11.9
Depth of body—
(a) Central		15		15.7		13.6
(b) At inner end of bow-cover		12.8		13		11.3
(c) At nner end of stern-cover		10.6		11.3		8.9

The canoes made on one island seem to preserve the same general proportions as regards the relative size and position of the various parts. The canoe-builders in each family group, however, have their own inherited style and any experienced tufunga can recognise at a glance certain peculiarities of workmanship which indicate the family by which a canoe has been produced. Most of the idiosyncrasies and small departures from the composite or ideal type which make possible this recognition are scarcely worthy of notice. There are two family traits, however, that merit description. One is the employment of a double after-boom as shown in Fig. 58. The additional boom is - 92 spaced some four to six inches forward from the true afterboom. It is usually identical in all respects with the booms already described, but sometimes it lacks a boom-leg (tapuvae).

The second is a longitudinal brace lashed on the outer side of the boom-legs and known as te taofi (See Fig. 46).

The majority of canoes, being made for use on the leeward side of the island, have not the double alignment of kalisi (Fig. 36), the use of which in a rough head sea will be explained in the succeeding section.

Again, the paucity of suitable trees is responsible for a number of freak shapes where the builders have had to follow the lines of badly-shaped logs. Canoes are not infrequently seen with the greater part of the bow or stern cut from another block of wood and laced to the hull. Such inserted pieces are termed tao.

About five years ago I saw, on Vaitupu, an outrigger of the type shown in Fig. 60. This canoe has long since broken up, and the sketch was drawn from a model made for me by the tufunga who built the original.

There is at present one canoe among the twenty or more which line the beach in front of the village, which has been cut from a log large enough to give the sides the requisite height without the addition of oa. This canoe has been executed in the traditional Vaitupu style. It is remarkable in two respects: the gunwales are half as thick again as those of the average canoe, and there are two thwarts, in one piece with the hull, in the neighbourhood of the forward and after-booms. These were left solid when the interior was hollowed out. This canoe is said to have been built about fourteen years ago, and is now the oldest canoe in use on the island.

Of somewhat greater importance are the differences in construction existing on different islands throughout the group. As was stated in the introduction to this paper, no detailed examination of the canoes in use on the other islands has been possible. Two points, however, in which canoes made on the northern islands of the group differ from those already described have been brought to my notice by an old tufunga on Vaitupu. One of these, the position of the outrigger platform which has already been noted, I remember - 93 having seen on my visits to the northern islands. For the other I have only the word of the tufunga. He claims that the Vaitupu arrangement of a short bow and a long stern, as figured throughout this paper, is sometimes reversed. In other words, some canoes, on other islands of the group, may be found to have the outrigger set much farther towards the stern than any canoe herein described.

Lastly, of latter years, several Nanumea type canoes have been built on Vaitupu, to which the bow and stern-covers are not permanently lashed in the traditional Northern Islands style, but are detachable as is the Vaitupu fashion, for the more efficient drying-out of the hull. A specimen pair of this new type of puke is to be seen in the background of the photograph in Fig. 69. The small projections at the front of the bow-cover and the rear of the stern-cover are to prevent the forward and after ends respectively from slipping sideways, since only the inner ends can be conveniently tied down. The lashings of the other covers in this photograph had to be cut in order to detach them from canoes.

It may fairly be assumed that whether the outrigger was developed from the double canoe or vice versa, it has reached its present form only after many years of cut-and-try experiment. And there is some attraction in setting forth the reasons for the various shapes and angles of adjustment in the probable order of their development.

The addition, always on the port side, be it noted, of an outrigger to a canoe for the sake of stability brought with it several disadvantages, the chief of which was an increased difficulty in swinging the canoe to starboard even from a stationary position. The farther the float was set from the hull the greater was this difficulty. On the other hand, the nearer the float was set to the hull, the less was the stability of the canoe. The result of this has been the setting of the float at the minimum lateral distance for safety in the open sea. Reference to the tables given above - 94 will show that this distance is about three quarters of a fathom (fatuli). 24

The difficulty of steering to starboard was found to be lessened further by setting the forward end of the float in towards the hull in such a manner that the medial lines of both hull and float would, if produced forward, intersect (See Fig. 36). The greater the angle at which the float is set, the easier it becomes to swing the bow to starboard. Here again, however, there was a limiting objection. It was found that setting the float at an angle to the hull increased the resistance of the outrigger during the forward motion of the canoe. The result of this has been to limit the amount by which the forward end of the float is set to starboard to about one hand's breadth from a line through the after end of the float parallel with the hull.

The assymetrical shape of a well-built hull which serves to counteract the tendency of the outrigger-float to steer the canoe to port while moving on a straight course has already been noted (Figs. 36 and 48). It was found, however, that when the canoe was being paddled against a head wind and sea, the thrust of successive waves against the more rounded longitudinal arc of the port bow threw the bow to starboard much more than was necessary to balance the pull of the outrigger-float. Thus the action of a canoe moving against the sea was extremely erratic, and the steering consequently difficult.

The problem then was to find a method of shaping the hull which would preserve its outrigger-balancing property and at the same time resist any increase in this function caused by the action of a head sea.

The ingenious system of setting the V-shaped keels (kalisi) at the bow and stern on different but parallel alignments (Fig. 36) was the method by which this difficulty was - 95 met. It will readily be seen that the swift pitching motion imparted to the canoe by the action of a head sea causes the hull to have a seesaw motion with the rounded bottom amidships for a fulcrum, as it were. Thus the bow and stern have alternately a short sharp downward plunge through the water. There are therefore two directions from which the waves of a head sea cause a movement of water relative to the sides of the bow and stern. The first is the ordinary motion of the water towards the stern caused by the movement of the canoe, and accentuated by the thrust of the waves which, as has been shown above, tends to throw the bow to starboard. The second is the upward motion of the water relative to the bow and stern alternately, caused by the pitching of the canoe. Now with the V of the forward keel (kalisi) pointing, not vertically downward as would be expected, but set, and therefore pointing, somewhat to the port side, the downward plunge of the bow of the canoe in motion causes a lateral deflection to port. Similarly the set of the after keel (kalisi) to starboard of the longitudinal medial line of the hull causes the stern to be pushed to starboard by the upward thrust of the water. Here then we have forces acting on both bow and stern which tend to steer the canoe to port and therefore counteract the tendency of a head sea to thrust the bow to starboard.

In calm water these keels have no steering-action since they are set parallel to the longitudinal medial line of the hull; while with a following sea there is little of the seesaw motion and certainly no swift upward thrust of water which would cause their steering-properties to operate.

In other times, each canoe had its proper name, and in some songs recording the prowess of a particular captain and his crew only the name of the canoe may be expressed.

• Te Ikalelau (a canoe named after a legendary Gilbert Islands warrior) went out
• Looking for te Pua Foua (another canoe lost in bad weather).

- 96

• It is lying inactive there on the ocean (i.e. the crew are not paddling)
• Waiting for the shoal of bonito (to show in what direction it is moving that they may follow)
• Because (they think perhaps) it is only shoal of kamai (another fish).

The following two lines are the beginning of a very ancient canoe-building song, the rest of which is not known to the present generation.

These references may, or may not, indicate that the canoe was held in greater reverence in former times than is the case to-day. Certainly one would expect that, when the building of a canoe was a laborious operation occupying many days of fatiguing work with soft shell adzes, some care would have been taken in its preservation. It is therefore strange to find that the only protection against the elements provided for the canoe while it was lying ashore was a covering of coconut leaves (taomanga) bound in twos or threes (taolua or taotolu) and thrown lengthwise over the hull, sometimes weighted down by a few large stones or pieces of wood.

The average canoe is in use almost every day. It may be uncovered (suke) as early as 4.30 a.m. and taken many miles out to sea to be on hand at the bonito grounds when the rising sun brings flocks of hungry noddies (ngongo and lakia) swarming out after the shoals of tiny fish (kavalikiand o) on which the bonito feed. If the bonito catch is poor, the canoe may proceed to the pala and other fishing grounds and remain occupied in snaring pala or in deep-line fishing for the albacore (takua) until late afternoon, when it will return to the village. At evening again it may be required to engage in the communal herding of flying-fish beyond the edge of the outer reef just as the sun is setting and when the flying-fish are packed in shoals unable either to fly, or to swim more than a few feet below the surface.

Then at the dark of the moon several hours may be spent each night torch-fishing for flying-fish and later a palu (Ruvettus) fishing expedition may at times be organised, and the canoe will not get back to the lango (the two small - 97 transversely placed logs upon which it rests when not in use) until sunrise.

It thus may be that a canoe is used occasionally with changes of crew for almost twenty-four hours in the day.

At the end of a fishing expedition, after the canoe has “shot” the breakers of the outer reef and has been carried, or at high tide been paddled through the comparatively tranquil shallow water on the reef up to the village beach, it is reversed and the stern is grounded on the sand of the beach. The “catch” is then carried ashore, the canoe bailed out, and carried, stern first, to its lango on the sand some few feet above high water mark. Here the bow and stern-covers (on the Vaitupu type canoe) are removed and the interior is thoroughly dried out. The taomanga is then thrown on to the hull and paddles, bailer and all fishing gear are taken away by their respective owners.

On a day unsuitable for fishing the whole fleet (on Vaitupu some 20 to 30 canoes) may be seen lying on their lango, covered by taomanga, bows pointing to the sea. This is a favourite playground for the young folk of the village, but all canoes are protected by a strictly enforced tapu.

1. An ordinary hole, stove in the side, if of small dimensions is repaired by plugging with a piece of puka wood cut to fit tightly and exactly.

A crack or split in the hull is repaired thus:—A sliver (tuatika) of the rounded back of a dry coconut-midrib (lafo) is split off (isi). This has a width of about .8 in. It is placed, rounded surface upwards, lengthwise, over the crack inside the hull, and holes are bored through the hull in pairs, one on either side of the tuatika along its periphery at intervals of one span. The sides of the crack are then reeved together by several turns of fine uka-lau-f ou (line made from hibiscus bark) through each pair of holes and over the tuatika. The holes, which are of very small diameter are afterwards plugged with fou (hibiscus) wood. This wood is chosen because it has the property of breaking off neatly and easily, flush with the surface of the hull. This operation is known as fakatuatika

- 98

• taumua, ahead.
• mata-ama, ahead of the outrigger (i.e. port bow).
• mata-katea, starboard bow.
• katea, abeam—on the starboard beam.
• tua-ama, abeam—on the port beam (lit. beyond the outrigger-float).
• te alovasa, beneath the outrigger.
• mulivaka, astern.
• muli ama, abaft the outrigger-float (i.e. the port quarter).
• muli-nofanga, the starboard quarter.

• afea, when?
• alo, upper surface of log, front of body.
• ama, outrigger-float.
• anga, span.
• asu, bailer.
• atu, bonito.
• atu-pa, adze with curved, laterally-reversible blade.

• faka to ifo, to "let in"; mortice.
• fakafiti, clamp made of a loop of cord and two sticks.
• fakatipa, Northern Ellice dialect for atu-pa.
• fangainga-o-lupe, about 9 a.m.
• fatinga-a-maikao, a unit of measurement; a finger-joint.
• fatuli, a unit of measurement. Length from tips of extended fingers of one arm to flexed elbow-joint of the other.
• fetau, Callophyllum inophyllum.
• foe, paddle.
• foe-uli, steering-paddle.
• foliki, small.
• fono, inserted piece in side of canoe.
• fou, hibiscus.
• foulua, double canoe.
• foutake, ancient method of lacing planks.

• inaki, the bulge on the side of a canoe.

• kainaki-tutonu, mid-day or afternoon meal.
• kai-o-moe, evening meal.
• kalisi, keel having a V-edge.
• kanava, Cordia subcordata.
• katea, starboard.
• kaufuatanga, long hand-grip superimposed on the sides of the Nanumea type canoe.
• kaupalepale, shelf of small sticks on the outrigger-booms.
• kautaka, projecting bulge of the side in the region of the oa seam.
• kau-tufunga, squad of canoe builders.
• kavaliki, tiny fish on which bonito feed.
• kiato, outrigger-booms.
• kofe, fishing-rod.
• kolokolo, two-ply twisted sennit.

• lakia, noddy (Micranous leucocapillus).
• lango: (1) Transverse rests placed under canoe and float fore and aft to raise them off the ground.
  (2) Piece superimposed on boom under the kaupalepale.
• lango-kofe: (1) Forked rests for carrying the bonito-rod on the outrigger when not in use.
  (2) The cross-piece on the Nanumea type canoe used as a rest for the bonito-rod.
• li, the lashing fastening the saunga on the hull to a projecting crosspiece from the gunwale.
• liu, the interior of the hull.
• lotu-a-toki, a religious ceremonial performed over the canoe-building adzes before the commencement of the day's work.
• lua, two.
• lualua, double canoe.

• mafuli, to capsize.
• maikao, finger.
• manga-lua, forked.
• manu: (1) v.i., see sumu. (2) Lifting-grip lashed across gunwales.

- 99

• mata-pili, a sennit grummet in which the butt of the bonito-rod is placed in order that the lure may troll while the captain is paddling.
• milo, Thespasia populnea.
• mua, fore.
• muli, aft.

• ngafa, fathom.
• ngasu, Soaevola Koenigii.
• ngie, Pemphis acidula.
• ngongo, noddy (Anous stolidus).
• ngutu, mouth, edge, gunwale.
• ngutu-puke, inner edge of (bow or stern) cover.
• nofoanga, seat.

• o, small fish on which bonito feed.
• oa, top-strake or plank superimposed to raise the side of the canoe.

• pāla, a large fish.
• palu, Ruvettus.
• palengalu, wave-guard, the sloping surface of a Nanumea style bow-cover.
• pono, peg driven into lacing-hole to clamp the lashing-cord.
• paopao, small canoe.
• pua, Guetarda speciosa.
• puka, Hernandia peltata.
• puke-mua, bow-cover.
• puke-tua, stern-cover.

• saunga: (1) Lifting-piece on outrigger.
  (2) Lifting-grip on port side of body, aft.
• siki, club.
• suke: (1) The lower of two raised platforms on the stern-cover (a rare occurrence— the upper is the tukungane).
  (2) v.t., to uncover (a canoe).
• suki, peg in outrigger-float to which boom-leg is attached.
• sumu, v.i. (of lashing) to have a diamond-shaped woven figure.

• tae, hand-net for catching flying-fish.
• ta-fa, adj., having four paddlers.
• tafola, whale, porpoise.
• taka-oa, long sliver of wood packing the lashings of the oa seam where they are made to stand out from the side by the kau-taka.
• takua, large kind of bonito.
• to-lima, adj. (of a canoe) having five paddlers.
• tao, a piece shaped from another log and reeved to the hull of a canoe to build up the bow or the stern.
• taofi, a brace lashed to the boom-legs.
• taomanga, covering of coconut-leaves bound in twos or threes and thrown on the hull when the canoe is lying ashore.
• tao-lua, (of coconut leaves used for covering a canoe) bound in twos.
• taotaoama, the stick forming the outer edge of the kaupalepale-shelf.
• tao-tolu, (of coconut leaves used for covering a canoe) bound in threes.
• tapalima, hand's breadth plus extended thumb.
• tapuvae, boom-leg.
• taumua, bow.
• taumuli, stern.
• tausunu, Taunefortia argentea.
• tautai, canoe-captain, steersman.
• tiale, Gardenia taitensis.
• tino-imua, bow-paddler.
• toki-fasua, shell adze (made from clam shell).
• toki-uli, stone adze (not obtained locally) (lit. black adze).
• tokoulu, the raised portion on the Vaitupu type stern-cover, containing a groove for the bonito-rod.
• tona, see suki.
• tua, exterior of hull; keel, amidships.
• tua-fa, having four turns of lashing-cord.
• tua-ono, having six turns of lashing cord.
• tua-tolu, having three turns of lashing-cord.
• tufunga, canoe-builder.
• tuki, see suki.
• tukungane, raised platform on stern-cover for bait.
• tuli-kaf a, three-plait sennit.
• tutonu, mid-day.

• vaka, canoe.

- xiii - xiv - xv - xvi - xvii - xviii - xix - xx - xxi - xxii - xxiii - xxiv - xxv - xxvi - xxvii - xxviii - xxix - xxx - xxxi - xxxii - xxxiii - xxxiv - xxxv - xxxvi - xxxvii - xxxviii

- xxxix Page is blank