Other formats

    TEI XML file   ePub eBook file  

Connect

    mail icontwitter iconBlogspot iconrss icon

Arts and Crafts of the Cook Islands

Fishing

Fishing

Social and Religious Importance of Fish

Fish were of the highest economic importance, providing a basic protein food to supplement vegetable foods, but they also had an important social significance. All social functions in which the people shared were accompanied by a feast. Before the day of the feast, a fleet of canoes went out to catch fish for the occasion. In Mangaia, fishing started days before the feast. The day's catch was cooked immediately in leaf wrappers, which helped to preserve it, and stored on raised platforms out of the way of rats. Each day the cooked fish were examined and if there was any sign of mould, the packages were recooked. The social prestige of the people giving the feast was raised by the quantity of food provided, and fish was a necessary part of the feast. Tapus on certain fish reserved them for chiefs. Men in an inferior position could curry favor with influential chiefs by taking them gifts of fish. Young men ingratiated themselves with girls by taking them presents of fish, and the acceptance of the fish was a sign of favor toward a lover's suit.

Fish had their tutelary deities, whose simple shrines were erected near the shore. The fisherman passing a shrine on his way to his canoe, dropped a piece of taro or even a pebble before it to promote success. On his return with a good catch, he offered a fish as a mark of gratitude. During ritual on the important temples, fish of large size and good quality were offered to the gods. The importance of fish as offerings in the earlier stages of ritual development is proved by the fact that when more elaborate ritual demanded human sacrifice, the human sacrifice was termed a fish (ika). In Mangaia there were two such "fish" sacrifices. One was the voluntary sacrifice of chief or priest of a defeated tribe to bring his tribe success in an impending campaign against its conquerors. He offered himself up to his tribal god and went out alone to be killed by the enemy. This form of sacrifice, termed ika tea (white fish), was for the purpose of "breaking the rule in power" (kia motu te 'au). The other form was used after a successful battle to install the new leader on the great temple of Rongo where he was given the power (mangaia) over the island. Hence, the sacrifice was termed the ika ta mangaia (the fish to give power). The sacrificial victims were taken from tribes (kopu) that had been defeated, thus such tribes were termed kopu ika, literally fish tribe, but idiomatically, the tribe that supplies human sacrifices. Such victims were also referred to as ei ika 'akatangi pa'u (fish for sounding the drum). After the victim was offered up on the temple, the peace drums were sounded.

page 209
Distribution of Fish

After community fishing with nets and leaf sweeps the catch was piled up on the beach and a distribution or sharing (tu'a) was made to the families that had engaged in the work. Women and children gathered on the outskirts to view the catch and the distribution. The head fisherman took the large fish of a similar kind and laid out one for each family. He then added fish in turn to each heap until the pile was exhausted. Chiefly families were given priority in distribution of the better or larger fish, and a family of greater numerical numbers might be given a few extra. The official distributor used his own judgment based on a thorough knowledge of the number and status of the various families. No squabbling took place at the distribution, but persons who felt that they had not received their just due might nurse a grudge that flared into acts of hostility. The customary method of distribution is still carried out, and it is characteristic of Polynesian hospitality to give shares to resident Europeans even though they may not assist in the fishing operations.

Rights of Chiefs

Certain fish, such as the shark and urua, were restricted (tapu) to chiefs. They were presented to the ariki of the district and he could give shares to the other chiefs. After complying with his chiefly obligations, he could return a sh'are to the fisherman. The mackerel (koperu) in Atiu and Mauke was a tapu fish restricted to the owners of the mackerel grounds near the reef. However, the owner could not only make presents of fish caught by his servants but he could give permission to another chief to take some fish direct from the grounds.

It was customary for fishermen to recognize the position of their chiefs by making them presents of choice fish. Some greedy chiefs imposed on fishermen by sitting on the path opposite the channel by which fishermen returned to shore. The fisherman had to make the best of an awkward situation by opening up his basket of fish, and allowing the chief to pick out the best fish. It is related of a Rarotongan chief that he carried out this plan so often that he became a terror to the local fishermen. One astute relative of the chief returned by a different channel. Safely home, he selected what he considered a fair offering and by a circuitous route approached the chief's house from the forest at the back. The chief received the offering somewhat sarcastically, saying, "I never knew before that sea fish were caught in the forest."

This chief's tyranny became so great that he took all the fishermen's catches. His method was to hold up one finger and say, "Give me a fish for this." When he exhausted the fingers on one hand, he commenced on the other hand. If there were any fish left when he had exhausted both hands, he started on his toes. The people prepared to rise against him, but one of his leading mataiapo chiefs conveyed a warning to him in a subtle manner. He asked the page 210chief to accompany him on a walk through the village to see his people; but as they approached each house, its door was closed against him. At the end of the village, the mataiapo said to him, "You see the feeling of the people against you. My advice is that you go before anything happens." That night the chief fled with his wife to another district.

Prohibitions

Some methods of fishing were called tautai tapu, because there were certain prohibitions regarding the conduct of the fisherman and his family connected with them. In Mauke, deep sea fishing for manga came under this category. Before leaving at night, the fisherman must not cohabit with his wife. His clothing and sleeping mat must not be handled by his children. His wife and children must not visit a neighbor's house, nor were other people allowed to enter his house. The children were not to quarrel or cry during the night of fishing. The person who had carried stones to his canoe must not make love on that night. Cooking fires must be put out immediately after use. Infringements of the rules were blamed if the fisherman caught no fish. Infringements ashore were revealed to the fisherman at sea by various signs. If his line caught on the bottom, evil had occurred in his house (kua kino tona ngutu'are). If the stone sinker was dropped three times while being attached, the person who had collected it had broken the rule about lovemaking. After such omens, bad luck (popoa) dogged the fisherman. If no ill omens occurred he had good luck (popo marie). Night trolling with the wooden hook was subject to similar restrictions.

Fisherman's Snack

There seems to have been no prohibition, as there was in New Zealand, about taking cooked food to the fishing grounds. Cooked taro was taken along to be eaten with fish on return to the beach. When the leader of the fishing fleet headed for shore, the rest all followed. The various channels through the reef had their hereditary owners, and when the canoes landed, the owner of the channel through which they had come in took a levy of fish from the canoes. In the owner's absence, his nearest of kin, took the levy. A fire termed a'i ika (fish fire) was lighted on the beach, the levy was cooked, and then distributed among all the fishermen regardless of whether they had been successful. After the meal, the lucky fisherman with his catch minus the small levy and the unlucky fisherman with at least a meal in his stomach returned home.

Fishing Technique

The accumulated experience of generations guided the technique adopted in the preparation of material for fishing (tautai) and the method of catching page 211fish. Native fishermen went to certain places at specific times to catch special fish by methods needed for that kind of fish.

Though all men were skilled, some were especially expert and well versed in fishing lore. Such men directed community operations in fishing drives within the lagoon and commanded the movements of the fishing fleet beyond the reef. Each man fished individually from his canoe, but a number of canoes usually went together to the fishing grounds. Experts knew the time to expect shoals of certain fish and young fry. They knew the right fishing grounds for the various seasons and the favorable nights of the lunar month, termed the ara po (night path). In Mangaia, the expert leaders were termed makona, and the knowledge they possessed was handed down in their families. Some fishermen were especially versed in the topography of the lagoon and reef and could find their way about in the darkest night without falling into holes and cutting themselves on the dangerous sharp edges of the rims formed by the bores of the ungakoa. These men were called tautai tonga tuatua akau (fisherman who knew the depths of the reef). Their knowledge was useful, for some fish were plentiful on dark nights or in stormy weather.

In community fishing, a scout was sent ahead to locate fish. He was termed mataika (mata, eye; ika, fish) and had, as the name implies, an eye for fish. He recognized patches of darker color or a ripple on the surface of the water that were made by shoals. Some informants say that experts could smell fish.

In Mangaia, the experts were said to be able to interpret what was known as the knonga. The kuonga was supposed to be indicated by the direction of the head of the Mango-roa (Long-shark), the native term for the Milky Way. It indicated the direction of the offshore grounds in which fish were plentiful. I could not get practical details, and I doubt whether my present day informants knew details.

Off-shore fishing was confined to angling with hook and line and the use of a scoop net on a long handle to catch flying fish. Canoes were necessary, and these methods were confined to men.

In-shore or lagoon fishing comprised a greater number of methods that were shared by women. Women collected shell fish and crustaceans, poked squids out of holes, and caught fish by hand groping and with nets. Fishing operations were conducted by day, and by night with the aid of torches. The people inhabiting a stretch of coast knew every rock and channel in their part of the lagoon. Net fishing depended on the state of the tide, most methods being carried out at low tide and on the turn when it came in. The Mangaian terms for the stages of the tide are as follows: low tide, tai va'a; full tide, tai pi; incoming tide, tai kapu; outgoing tide, tai 'eke.

The following chant refers to crustaceans caught on the beach in the territory of Kainuku, One of the ariki of Rarotonga.

page 212
Nunui 'ua te ika Abundant are the fish
i te ava o Kainuku; in the channel of Kainuku;
Ko te ko'iti-rau-kura, The red tinged sea crabs,
Ko te tupa-vaevae-roa, The long-legged land crabs,
Ko'ia ki te kikau, Gather them into the coconut leaf basket,
Ka pu'era 'ua te ka'i. The clams in their plenty open their valves.

Within the lagoon and on the reef, the methods of fishing included hand groping, treading, torching, poisoning, spearing, trapping with weir and trap, netting, and angling.

Hand Groping

Hand groping (naonao) is used by men and women, and they are skillful in catching fish by hand. Long experience has taught them the configuration of the various rocks and they grope in the holes and clefts for rock frequenting fish. People driving fish toward a net never pass likely looking rocks without groping for hidden fish.

Sometimes three or fo ur men with sticks drive pa'oro fish into a rocky pool, where the fish are procured by hand groping. This procedure is termed pakoti pa'oro because the pa'oro are cut off from escape (pakoti). It is dangerous to touch some fish—such as the sea eel (kara'oa, Aitutaki), poison fish (no'u), porcupine fish (totara), and echinoderms (vana)—hence some fishermen use a two foot blunt rod of ironwood to feel ('akapaki) in the holes and crevices. When a fish is found, the stick is run gently along it and the fisherman can tell from the feel of the spines and the shape of the fish whether he should grasp it.

Treading

Besides being caught by hand, in nets, and by angling, crayfish (koura) are caught by treading (taka'i) upon them as they crawl about on the bottom of the lagoon during the dark nights of the lunar month. Most suitable are the last night of the month and the first four nights of the new moon. Torches are used to give light, and when a crayfish is located, a foot is quickly placed upon it with enough pressure to keep it a prisoner. The fisherman bends down, slides his hand under his foot, and grasps the body of the crayfish when he removes his foot, he twists the crayfish belly upward to prevent its kicking and places it in the basket tied around his waist.

Torching

Torches (rama), made of dry coconut leaves, are used in night fishing within the lagoon and off shore in netting flying fish. The required number of leaves are placed one above the other with their butt ends together. The butt ends are tied together with a strip of hibiscus bark (70, p. 279, fig. 236); and page 213the operator, straddling the leaves, pushes the butt ends forward as he binds the strip tightly in spiral turns by rotating the bundle. The bark is tied at the tip end. Torches are qualified by the form of fishing in which they are used. Aitutaki people classified three forms (70, pp. 279, 280), as follows:

1.Rama tupa'i are made of three leaves which the fisherman carries in his left hand. He has a basket tied around his waist for the catch and carries a piece of curved ironwood, termed rakau tupa'i, in his right hand with which to strike fish attracted by the light of the torch. [Now a piece of hoop iron (pa'a) is used.] This method is used in the shallower water near the beach or the reef. While seeking fish, crayfish are also procured by treading.
2.Rama patia (patia, to pierce) are larger, consisting of four or five leaves. These are carried on canoes in the deeper waters of the lagoon and are held by a torch bearer (tangata turama), who stands behind a spearman in the bow while a third man poles the canoe. As one torch burns down, the torch bearer picks up a fresh torch which he lights at the tip from the burnt down torch before it is thrown overboard.
3.The rawia maroro (maroro, flying fish) torch is used outside the reef in a method similar to spearing, but the bowman is armed with a flying-fish net and the third man paddles. As the fish fly toward the light, the bowman intercepts them with the net. Fish in the water are also caught by a dexterous use of the net. The net method can also be used for other fish within the lagoon.

Torches may be made also from the dry flower spathes (roro) of the coconut, augmented with dry ironwood twigs (jara too), and are called rama roro. When dry kernels of the candlenut (tuitui) which are oily and give a bright light, are added to the rama roro, the torches are called rama tuti (tuti, a shortened form of tuitui). These two types of torches were formerly used within the lagoon, but, as they are harder to make, the coconut-leaf torch has supplanted them.

Octopus Catching

Octopus ('eke) is esteemed as food and as bait for angling at sea. The body of the octopus is termed the pu; the tentacles, kakave; the eyes, mata; and the beak, ni'o. The Cypraea- shell octopus lure used in various forms in the Society Islands, Hawaii, Samoa, and Tonga seems to be unknown in the Cook Islands, though I have a questionable reference that it was used in Rarotonga. Methods of catching octopus differ within the shallow lagoon and the deep water beyond the reef. The following methods apply to Mangaia.

Lagoon Fishing

Small octopuses within the lagoon are termed 'eke maori, and the process of catching them is termed taka'i maori. They are caught principally by women, who use a pointed stick (ko) of ngangie about 2.5 feet long. At low water, the women search for likely looking holes on the lagoon bed between the shore and the reef. The stick is thrust into a hole and stabbed about to drive out the octopus. As it emerges, its body is grasped in the woman's hand; and it is killed by driving the pointed stick between the eyes. (In Hawaii and Samoa, page 214the octopus is bitten between the eyes instead of being killed with a sharp, stick.) The dead octopus is placed in a coconut-leaf basket slung around the waist. Women still comb the bottoms of shallow lagoons for octopus.

Deep-Water Fishing

Deep diving methods are no longer carried out by men, but the following was the method used:

When the fishing canoes passed through the reef channel, a search was made for octopuses to provide bait. Diving for octopuses in the deep water outside the reef was termed pae 'eke. The successful fishermen must have good sight and a sound wind. Each man carried three sharpened sticks of ngangie wood 2.5 feet long to assist in driving the octopuses out of holes. They had. no fear of the tentacles fastening upon their arms once the octopus was within reach, for a firm grasp upon the body caused the tentacles to relax. However, when an arm was thrust into an occupied hole, the octopus seized the arm with its tentacles and pulled toward its body. If the arm was pulled back, the octopus pulled the harder and the fisherman was likely to be kept a prisoner until he drowned. The native fisherman, however, did not pull back, as he was anxious to get his hand on the body of the octopus. He accordingly pushed his arm farther, and Mangaians say that when the octopus found that the arm was drawn in without opposition, it let go. On locating an octopus on the bottom of the sea, the fisherman tied a line around one of his ankles after fastening the other end to a boom of his canoe to prevent it from drifting too far away. The Mangaians recognize the following five varieties of octopus, and give slightly different techniques for catching each.

1.'Eke para'i is the single octopus which has no hole in the rocks, but seems to wander about. The fisherman dived down, seized it by the body, and came up with it to his canoe, where he stabbed it with one of his sticks.
2.'Eke tau'ani (tau'ani, embrace) are in pairs, the smaller female being near a hole while the larger male is farther away. They touch or embrace each other with their tentacles. The fisherman took a stick down with him in case they escaped into the hole. He picked out the female and placed his hand by the hole so that when the female turned over to enter the hole, the body came toward his hand. Grasping the body firmly, he rose to the surface, killed the octopus, and placed it in the canoe. After a rest, he went down for the male.
3.'Eke taeva has a hole about a foot in diameter sufficiently large to allow the free insertion of the arm. If the octopus was in the hole, a hand was thrust in and worked back to grasp the body. The other hand might grasp the beak and the 'eke taeva was brought up.
4.

'Eke paepae is so called because it blocks its holes with stones and thus gives rise to the metaphor of the stone platform (paepae) of a house. Several dives had to be made to remove the stone obstructions, before the octopus could be caught. It is said that this octopus imight have pushed forward other stones after the first ones were removed. The fisherman removed the stones, and thrust a sharp stick into the body of the octopus where he left it as he came up for air. He went down with a second stick which was manipulated to force the octopus within reach. If three feelers protruded it was a good sign. The fisherman attempted to stab the octopus in a vulnerable part of its body. If he got short page 215of wind, he left the second stick in the octopus, and came up for air. Then he descended with the third stick and drove that into the body of the octopus. If a fourth feeler was protruded, it was a sign that the octopus was uncomfortable and would shortly vacate its hole. As it came out, the body was grasped.

During the different stages of the attack, the fisherman kept his eyes fixed on the bottom while resting by his canoe, for the octopus might vacate its hole. With any movement below, the fisherman dove down immediately to follow up the octopus. If the octopus squirted out black fluid, the fisherman avoided it by diving quickly to the side.

5.'Eke tapairu are very large black octopuses. They were tackled by daring fishermen in the olden days, but sometimes a man lost his life.

Fishermen who were successful in catching octopus for bait were surrounded by the fishing fleet, and portions of the feelers were shared. In his youth my informant, Aiteina, had carried out the various methods described above.

Fish Poisoning

Fish poisoning ('ova) is carried out in suitable pools (roroka) in the lagoon. Such pools have distinctive place names. The narcotic is obtained from the grated kernel of the fruit of the 'utu and from the crushed leaves, stems, and roots of the mata'ora. The fruit of the 'utu is mature in September but the mata'ora plant is available at any time. If the 'utu fruit is gathered before time to use it, it must be covered to prevent it drying. The fruit is split to extract the large kernel which is grated and placed in rough coconut-leaf baskets. To poison the pool, the operator wades through it carrying the basket of grated material below the water surface and shaking it as he goes to distribute the material through the interstices between the wefts of the basket. The mata'ora is spread by pulling handfuls out of the basket, and distributing them evenly, and diving down to place handfuls in rock crevices to drive the fish out. Larger fish are narcotized, and they move about sluggishly so that they are easily speared or caught in scoop nets. The young fish die, and many are lost because they remain in the rock crevices.

Small pools may be poisoned by a few people, but sometimes large scale fish poisoning is a social function to which various villages are invited by the host village which owns the pools to be used. I took part in a large fish poisoning in Rarotonga near the village of Arorangi in 1926, and the whole population of the island participated. The time had to coincide with low tide just before dawn. The Rarotongans held that after feeding, the fish slept during the night but awoke with the dawn. When they swam down the channels toward the sea, their way was cut off by the reef which was above water at low tide and any open channels were shut off with nets. The beach along the lengthy pool was divided into sections for each village. The visiting villagers came during the preceding evening and camped amid the ironwood trees facing their section of beach. They brought baskets of grated poison for their section of the pool. In the early morning, the Women's Committee of Aro-page 216rangi , who conducted operations sounded a wooden gong and criers went along the beach calling upon the people to bring out their poison. The baskets of poison were then placed along the beach, each village in its own section. If inspection showed that some groups did not have enough poison, their stock was reinforced from local reserve supplies. Finally, the gong was sounded as a signal to commence poisoning the pool. Each man carried a basket in each hand and waded from beach to reef in a direct line shaking out the poison as they went. In this manner the whole pool was adequately poisoned. At least half an hour was allowed for the poison to take effect. Meanwhile, men, women, and children waited on the beach with baskets, spears, and scoop nets. At another signal from the gong, all dashed into the water to procure all the fish they could get. Once in the water, people could go where they wished, and there was splashing, laughing, and general enjoyment. The people took the fish to their camps beside the shore, cooked fish, and feasted. Before the feast however, a levy of fish was made on those who were most successful, and distributed among the sick, the aged, and the Europeans, who were somehow classed with the helpless. The full details of the method have been described elsewhere (71). The two tides after the poisoning threw up large quantities of young fish that had been killed. As the natives are aware of the waste that accompanies the 'ora method, it is used at long intervals, as much for social intercourse as to obtain food.

Some of the Rarotongan native missionaries brought back a plant from New Guinea which has been termed taru Papua (Papuan weed). The crushed root is used for narcotizing small pools, but the older Rarotongans condemn it because it is used too readily by individuals with a resulting waste of young fish.

I quote the following summary (71, p. 66):

The hora['ora] method of fishing depends primarily on suitable pools of large extent being left isolated within the reef at low tide. In other islands of the Cook group [besides Rarotonga] large pools are not available. In Aitutaki, there is deep water within the lagoon at low tide, and fish-poisoning does not exist. In Mangaia, and some of the other islands, smaller pools admit of poisoning on a small scale. Aitutaki furnishes the useful lesson that in remarking on the absence of such a culture-trait, the presence of the Barringtonia is useless without the physical conditions being suitable. For the Cook Islands, the [large scale] hora method, as described, is confined to Rarotonga. Hawaii has the same name for fish-poisoning in the form of hola.

Spearing

In pre-European times the fish spear was made from a straight piece of ironwood with one end pointed and without a barb. Spearing (patia) with iron-pointed spears is now a favorite method of catching fish and has lessened the use of other fishing methods. Spear heads with three or more barbed points are obtained from the blacksmith or trade stores and lashed to a long page 217pole of hibiscus. These are sometimes thrown with great dexterity, the fisherman not only judging distance but allowing for refraction in the water. A short spear with a single iron point without a barb is also used around rocks and is inserted into crevices in the rocks. The left hand is slipped along the spear point to hold the fish on, as the barbless point would slide away with a direct pull. Some men have a hook lashed to a stick which they use as a gaff to prevent the fish from slipping off.

Fish Weirs

Fish weirs (pa) are stone walls (pa) erected usually in the form of a V leading into an enclosure. They are set on the course of a channel within the lagoon with the open part of the race facing the shore. Fish going out on a falling tide are intercepted and pass down the converging arms of the V into the opening to the enclosure. Fishermen on either side of the channel watch the fish passing by, and when they enter the enclosure, a net is drawn across the opening. The fish within the enclosure are then speared or caught in scoop nets.

The sites of most of the weirs still in use were laid down generations ago, and the walls of coral stone usually need repairing from time to time. The weirs belong to the family whose ancestors built them originally; and the use of a weir without permission is regarded as theft, and formerly led to bloodshed. Aitutaki still has a number of weirs in use and they are classified as follows (70, pp. 298-303):

1.Pa kiokio is so called because one of the main fish caught is the kiokio. It is a V-shaped weir with an enclosure to the side. The length of the V-walls is lengthened toward the shore by lengths of coconut leaves attached to a vine (fig. 138, a).
2.Pa tute is temporary, consisting of two straight walls converging into the middle of a channel with an opening about a foot wide. There is no enclosure, but a large scoop net is placed at the opening to intercept the fish (fig. 138, b). A variation is formed by bending the narrow part at an angle so that the fish cannot see the net until they are practically in it.
3.Pa tuakirua is built on the same plan as the pa tute, but the wide inlet faces toward the sea. The apex of the V-walls is closed and covered over with stones to form a stone house ('are po'atu). The fish pass up the race and stay in the stone house. A fisherman stationed at a point midway along the walls closes the race with a large scoop net after the fish pass up. The fish are then driven from the house into the net with a pole (fig. 138, c). The commonest fish caught is the tuakirua which gives its name to the weir.
4.Arani is the name of the most famous weir in Aitutaki and the proper name indicates a distinct type. It is a rectangular enclosure set in a channel with an opening in the middle of the shoreward wall. There is no stone race, but its place is taken by diverging arms formed of coconut leaflet sweeps (ran). After fish pass into the enclosure, a net is drawn across the opening. Large fish are caught and the most prized are the ava (fig. 138, d).

Weirs varied in form but the four listed above comprise the main types.

The only other island where I saw weirs was in Rarotonga at Ngatangiia, where there were quite a number of enclosures that resembled a maze.

page 218
Figure 138.—Fish weirs, Aitutaki. a, pa kiokio, named Taketake: 1, 1, long side walls (kavi'i); 2, short end wall (po'o); 3, inner wall (tumu arero) to complete V-race and enclosure (4) with entrance opening (5); 6, rounded end (putaka); 7, 7, leaf sweeps extending V-shaped race, b, pa tute:1, 1, side walls; 2, opening where net is placed, c, pa tuakirua:1, 1, side walls converging toward shore; 2, stone house where fish take refuge, d, Arani fish weir: 1, 1, side walls here termed rauroa;2, end wall (po'o); 3, front walls (tumu arero) leaving middle opening (4) termed ara ika (fish path); 5, 5, leaf sweeps.

Figure 138.—Fish weirs, Aitutaki. a, pa kiokio, named Taketake: 1, 1, long side walls (kavi'i); 2, short end wall (po'o); 3, inner wall (tumu arero) to complete V-race and enclosure (4) with entrance opening (5); 6, rounded end (putaka); 7, 7, leaf sweeps extending V-shaped race, b, pa tute:1, 1, side walls; 2, opening where net is placed, c, pa tuakirua:1, 1, side walls converging toward shore; 2, stone house where fish take refuge, d, Arani fish weir: 1, 1, side walls here termed rauroa;2, end wall (po'o); 3, front walls (tumu arero) leaving middle opening (4) termed ara ika (fish path); 5, 5, leaf sweeps.

Fish Traps

The fish traps seen in the Cook Islands were of three forms: a circular, flat-bottomed trap with an opening at the top like a lobster pot, a cylindrical trap with openings at either end, and an eel trap with a funnel opening at one end. Two were in use in Aitutaki and the eel trap was in use in Atiu.

The circular traps ('inaki; Aitutaki, 'ina'i) in Aitutaki are made of the long thin roots of the coconut (aka niu). In Mauke, they are made of the pirita vine, and in Rarotonga of the aerial roots of the kiekie. The process of making a trap is termed raranga 'ina'i in Aitutaki, but the twining technique used is termed taviri in Rarotonga. The longitudinal elements or warps ('ora) are in pairs and the two twining elements or wefts (pokai) pass around the warps in a single-pair twine. The twining commences at the inner end of the funnel; and when the length is completed, the warps are bent over to form the body of the trap. As the diameter increases, additional warps are added and the weft twine is carried on in a continuous spiral. At the bottom, the warps are bent in at a right angle. The details of technique resemble that of the crayfish traps of Samoa (73, pp. 450-454). On the flat bottom' of the trap a cross piece of ironwood is lashed and four stone sinkers are lashed to its ends with hibiscus bark. Near the opening on the upper surface, a piece of wood is passed under three or more warps on either side to which short pieces of rope were tied and spliced together in a loop. When the trap is set in deep water a long rope is tied to the top to permit the trap to rest on the bottom, and floats (uto or poito) of breadfruit wood are tied to the upper end. The page 219trap is baited with fish, usually maito, which is tied to the inner surface of the trap near the opening (70, pp. 311-313). An opening with a lid cover (papani) is made for removing the catch, which consists of sea eels, kara'o, va'aroa, parakava, kopeti, tu'ua, teatea, koiro, and kakakakakuru.

In Mauke, the trap is termed 'inaki tuku pata because the sea eel (pata) is the principal catch. There are a number of floats of breadfruit wood, and the last float is a gourd ('ue). The bait used is koperu, and it is held that when a fisherman catches an extra quantity and eats some of them, the koperu bait used in the trap will not attract the sea eels. (See fig. 139, a and pl. 11, B.)

Figure 139.—Fish traps, a, section of 'ina'i trap: 1, funnel entrance (i'o); 2, cross pieces (te'a) on bottom with stone sinkers lashed to ends; 3, wooden supports for ropes; 4, two ropes joined in loop, b, technique of 'ina'i trap: 1, double warps ('ora); 2, single pair twined weft (pokai). c, section of anga trap: length, 54 inches; middle depth, 26 inches; end depths, 22 inches: 1, funnel entrance (matatapua); 2, central passage (i'o); 3, middle partition (paruru); 4, entrance (ngutupa) from passage to body of trap, d, warp and weft technique of anga trap seen from outside; continuous spiral hoop (1) passes below thick warps (2) and above thin warps (3); weft lashing element (4) consisting of a coir two-ply cord, passes obliquely across thick warps but makes a complete turn around thin warps. e, inside view of d.

Figure 139.—Fish traps, a, section of 'ina'i trap: 1, funnel entrance (i'o); 2, cross pieces (te'a) on bottom with stone sinkers lashed to ends; 3, wooden supports for ropes; 4, two ropes joined in loop, b, technique of 'ina'i trap: 1, double warps ('ora); 2, single pair twined weft (pokai). c, section of anga trap: length, 54 inches; middle depth, 26 inches; end depths, 22 inches: 1, funnel entrance (matatapua); 2, central passage (i'o); 3, middle partition (paruru); 4, entrance (ngutupa) from passage to body of trap, d, warp and weft technique of anga trap seen from outside; continuous spiral hoop (1) passes below thick warps (2) and above thin warps (3); weft lashing element (4) consisting of a coir two-ply cord, passes obliquely across thick warps but makes a complete turn around thin warps. e, inside view of d.

In Rarotonga, small traps made of kiekie in similar shape and technique are called tukutuku. They are baited with hermit crabs or fish and set in the lagoon with rocks piled around them to keep them in position. Such traps are without sinkers and without the long rope with floats. The fish caught are pakirikiri (rock cod) and pakoukou. Another trap of similar shape is made in Rarotonga from a creeping plant (ponu) that grows on the beach. These traps are only good during the week that the natural green color is retained.

The cylindrical trap (anga) in Aitutaki is made of rods of ironwood, thicker rods forming the longitudinal warps ('ora) with transverse elements (pokai) consisting of a number of twisted thinner branches crossing the warps and lashed together by a coir two-ply cord at their crossings. The trap is started with a narrow cylindrical passage (i'o) divided in the middle with a partition (paruru) with a hole (ngutupa, doorway) on either side. The passage page 220broadens out in funnel fashion, and then the outer cylindrical cover is formed to connect the expanded ends of the funnel entrances. The details of technique are described elsewhere (70, pp. 308-311). The technique resembles somewhat the bamboo double entrance trap of Samoa (73, pp. 463-466).

The trap is baited and set in a channel with its long axis in the direction of the channel. It is weighted down with stones carefully arranged to look like a natural hole in the rocks. The fish enter the funnel entrance, pass into the narrow cylindrical passage till they reach the partition, and then pass through the opening into the body of the trap. Parrot fish (u'u) and rock frequenting fish are caught in this trap (pl. 11, A; fig. 139, c).

In Rarotonga, similar traps are made from the aerial roots of the kiekie (70, p.311).

Eel traps, also termed 'inaki, are made of a fairly thick vine from 5 to 8 mm. in diameter. The one described below was procured in Atiu. This trap has a wide end with a self-acting funnel, the other end is smaller with an opening over which a piece of coconut-husk cover is lashed to two parallel sticks pierced through below the rim. The four ends project outward and the cord or sennit is crossed diagonally over the cover and around the stick ends. The warps or rods and the two element wefts are of the same material, and the paired weft is turned around each warp in a half turn termed a single-pair twine. The twining commences at the end of the funnel, works out as an even cylinder, and then expands in two stages by the addition of more warps. At the wide end, the warps are bent over to form the body of the trap and after narrowing toward the other end, the open rim is finished off neatly. The trap is well made. (See plate 11, C, D, and for technique see figure 140.)

After the trap is baited through the small opening and the cover is attached, it is laid on its side in the same manner as the New Zealand eel traps (hinaki).

Leaf Sweeps

Leaf sweeps (rau) are used in the Cook Islands and elsewhere throughout Polynesia (70, pp. 277, 278). They are made of coconut leaves split down the midrib, and the extra butt thickness is split off and discarded. In Aitutaki, three half leaves are joined together by knotting the midrib ends and twisting the leaves so that the leaflets stick out in all directions, and these twisted leaves are tied to a long stretch of stout vine (kaka-vai). The whole sweep is termed miakaatu as well as rau (leaf). The ends are termed kaoti (finish) and the middle is called kopu (belly). The easily made sweeps are used in place of a net to drive fish toward a beach, or used as stationary arms to direct fish into weirs. Though fish can break through, they avoid the moving leaflets and are driven before the sweep. The whole village helps in hauling a leaf sweep.

page 221
Figure 140.—Atiu eel trap (Bishop Mus., C2850). a, mesial section of trap: total length, 495 mm.; greatest diameter, 255 mm.; diameter at small end, 101 mm.; length of funnel, 310 mm.; funnel at opening about 65 mm. (2.5 inches) in diameter, extends as an even cylinder (1) for about 158 mm.; increases in diameter (2) by addition of four more warps for about 72 mm., and still more (3) by addition of eight more warps for about 82 mm. which brings it to outer circumference (4); warps are bent over to form body (5) and later diminish in size to reach rim opening (6) at small end. b, commencement of funnel end opening with warps flattened out to show twining technique; six warps (1-6) are used but in order to prevent slipping over ends of individual warps, they are arranged in three pairs formed by doubling three pieces of vine so that bends will form free edge of funnel; weft elements also are locked by doubling one piece of vine (7) around first warp (1) and then its two elements around second part (2) of pair in a twine which is continued around other warps as shown; twine reaches first warp (1) and continues a second row close to previous one; continuous twine is continued over the six warps until first even part of cylinder (a, 1) is completed. c, diagram of open end of trap, showing addition of warps forming three areas eccentric to funnel hole in center: inner area (1) has six warps forming even cylinder (a, 1); intermediate area (2) adds four more warps to make a total of 10 in expanding part corresponding to a, 2; outer area (3) adds eight more warps to total 18 in area corresponding to a, 3; this brings expanding opening up to outer margin of large end (a, 4), but another one (4) is added to make a full complement of 19 which are turned back to make outer wall (a, 5); throughout, single-pair twine is carried out in a continuous encircling movement. d, addition of new warps: old warps (1-3) are shown covered at lower end by three rows (6) of twining; additional rods (4, 5) have been pushed down beside two old rods (1, 2) between pairs of three rows of twining; they were then diverged, and next weft row (7) took turns around each of the five warps as did the succeeding rows above them; after the first couple of rows, warps were kept parallel and the twining became even. e, addition of new weft elements: one element of a weft pair was always cut off shorter than the other so that additions could be made singly; in upper row formed by weft pair (5, 6) one of pairs (5) has passed around warps 1 and 2 but is too short to go past warp 3; a new weft strip (7) is pushed down beside warp 2 between pairs of three rows, old weft 5 is dropped on near side of warp 3, and new weft 7 takes its place to pass in front of warp 3 and behind next warp (4) to continue twining while other original member of pair (6) continues until so short it is replaced; the continuous weft twine is rendered possible by such replacements. f, rim finish: when body of trap is diminished gradually in size toward small end, seven of full quota of 19 warps are inclined toward neighboring warps that will maintain even spacing, and two such warps just before and after meeting are included under same turn of twine, and after being so fixed by a number of rows, one of each pair is dropped or cut off; 12 warps at small end; when second last row (1) of twining is being made as always from left to right, warp head (4) is bent over in a curve, laid on near warp (3), and included under turn of twine (1); similarly, next warp (5) is bent over on near warp (4) and caught under turn of twine of second last row (1); when circuit is completed, last weft row (2) merely repeats including bent over warp ends with warps on which they rest; on completing second and final circuit, weft ends are cut off after passing through arch made by last warp; two of 12 warps were cut off above last twined row without being bent over.

Figure 140.—Atiu eel trap (Bishop Mus., C2850). a, mesial section of trap: total length, 495 mm.; greatest diameter, 255 mm.; diameter at small end, 101 mm.; length of funnel, 310 mm.; funnel at opening about 65 mm. (2.5 inches) in diameter, extends as an even cylinder (1) for about 158 mm.; increases in diameter (2) by addition of four more warps for about 72 mm., and still more (3) by addition of eight more warps for about 82 mm. which brings it to outer circumference (4); warps are bent over to form body (5) and later diminish in size to reach rim opening (6) at small end. b, commencement of funnel end opening with warps flattened out to show twining technique; six warps (1-6) are used but in order to prevent slipping over ends of individual warps, they are arranged in three pairs formed by doubling three pieces of vine so that bends will form free edge of funnel; weft elements also are locked by doubling one piece of vine (7) around first warp (1) and then its two elements around second part (2) of pair in a twine which is continued around other warps as shown; twine reaches first warp (1) and continues a second row close to previous one; continuous twine is continued over the six warps until first even part of cylinder (a, 1) is completed. c, diagram of open end of trap, showing addition of warps forming three areas eccentric to funnel hole in center: inner area (1) has six warps forming even cylinder (a, 1); intermediate area (2) adds four more warps to make a total of 10 in expanding part corresponding to a, 2; outer area (3) adds eight more warps to total 18 in area corresponding to a, 3; this brings expanding opening up to outer margin of large end (a, 4), but another one (4) is added to make a full complement of 19 which are turned back to make outer wall (a, 5); throughout, single-pair twine is carried out in a continuous encircling movement. d, addition of new warps: old warps (1-3) are shown covered at lower end by three rows (6) of twining; additional rods (4, 5) have been pushed down beside two old rods (1, 2) between pairs of three rows of twining; they were then diverged, and next weft row (7) took turns around each of the five warps as did the succeeding rows above them; after the first couple of rows, warps were kept parallel and the twining became even. e, addition of new weft elements: one element of a weft pair was always cut off shorter than the other so that additions could be made singly; in upper row formed by weft pair (5, 6) one of pairs (5) has passed around warps 1 and 2 but is too short to go past warp 3; a new weft strip (7) is pushed down beside warp 2 between pairs of three rows, old weft 5 is dropped on near side of warp 3, and new weft 7 takes its place to pass in front of warp 3 and behind next warp (4) to continue twining while other original member of pair (6) continues until so short it is replaced; the continuous weft twine is rendered possible by such replacements. f, rim finish: when body of trap is diminished gradually in size toward small end, seven of full quota of 19 warps are inclined toward neighboring warps that will maintain even spacing, and two such warps just before and after meeting are included under same turn of twine, and after being so fixed by a number of rows, one of each pair is dropped or cut off; 12 warps at small end; when second last row (1) of twining is being made as always from left to right, warp head (4) is bent over in a curve, laid on near warp (3), and included under turn of twine (1); similarly, next warp (5) is bent over on near warp (4) and caught under turn of twine of second last row (1); when circuit is completed, last weft row (2) merely repeats including bent over warp ends with warps on which they rest; on completing second and final circuit, weft ends are cut off after passing through arch made by last warp; two of 12 warps were cut off above last twined row without being bent over.

page 222

The long makaatu is drawn out parallel with the shore when the tide is flowing in. Just before the ebb, the ends are drawn toward the shore with men following behind the sweep to splice breaks or lift it over rocks. As the ends reach the shore, they are doubled back to reinforce the part still in the water. In this way, the belly of the sweep is brought to land and the fish procured. The makaatu is used only once.

The torau is a short sweep, from six to 15 yards in length. The leaflets, which are split on either side of the leaflet midribs, hang down more than do those on the makaatu. The sweep is used by young people along the beach at high tide and may be used several times.

As accessories to weirs, the sweeps provide a V-shaped race for weirs of the Arani type which have no stone walls leading to the opening of the rectangular enclosure. Though the sweep is stationary, the leaflets move with the tide or current and the fish move on to the weir opening.

In Rarotonga, I saw the leaf sweep being used but the middle part consisted of a meshed net of cord. It was used near the outer reef because the meshed net caught the fish, whereas the complete leaf sweep had to operate near the shore to provide a dry place to drive the fish.

Netting

Nets (kupenga), which are used extensively throughout the group, differ in form to suit various methods of fishing. The unevenness of the lagoon bed, studded with holes, rocks, and coral upgrowths, is not conducive to the hauling of long seine nets; and it was probably this factor and the ease of manufacture that led to the use of leaf sweeps which did not tangle on the bottom. Yet long nets are used in selected places, mostly as stationary barriers into which fish are driven. The ends are curved around to form an enclosure to imprison the fish, which are finally speared or caught in scoop nets. Shorter nets attached to end poles are popular, they are also stationary and the fish are driven into them. Scoop nets attached to frames of varying size, with or without handles, are much used either in the open water or in the channels along which fish pass from and to the sea. All net fishing depends on the state of the tide. The fishermen move toward the shore with an incoming tide, as they cannot operate in deep water.

The short, finely meshed nets with end poles (nariki) were valuable property, and some of them became family heirlooms. Gill (30, p. 211) records a story of a fugitive of a conquered Mangaian tribe who was able to buy protection with such a family heirloom. Fugitives hiding in the makatea raised reef of Mangaia were able to obtain bark and make nets as a ransom to obtain liberty.

page 223
Net Making

Net making (ta kupenga) is essentially a male occupation. The cord now used is bought from the trade stores, but the old cord was made from the bark of the oronga, hibiscus, papako, or sometimes the paper mulberry. The oronga, which is the strongest, was identified by Gill (30, p. 9) as Urtica argentea, but the modern identification is Pipturus velutinus. Gill says that the natives stripped off a quantity of oronga bark and, after carefully scraping it, exposed it to the sun by day and to dew by night until it was ready for use. Mauke informants said that hibiscus bark was stripped from young trees and soaked for several days in the clear water found in caves (vai ana) to make the material white (teatea). If soaked in muddy water (vai vari) such as that of taro swamps, the material was brown. After three weeks, the outer bark (pakiri) was easily removed from the inner bast (kiko) by pulling the strips through the hand (perekuku) without using a shell scraper. It was dried, separated into strands, and rolled ('iro) on the bare thigh into two-and three-ply cords.

Netting needles were used, except in Aitutaki; and mesh gauges were used in all the islands. Netting needles or shuttles ('ika) were said to be of the usual Polynesian form such as those of Tahiti (fig. 141, a), which differ from the European needle (fig. 141, c). In Mangaia, a pointed stick was used to which loops of cord were fastened by clove-hitches (fig. 141, b). The Mangaians termed their needle kauta. The Mangaian type was also used in Mangareva
Figure 141.—Netting needles and gauges. a, Tahitian netting needle collected by G. Bennet before 1829 (British Mus., 7052): length, 417 mm.; middle width, 40 mm.; length of opening (1, 2), 44 and 52 mm.; length of slits (3, 4), 40 and 31 mm.; thickness of side edges, 26 mm.; middle thin portion (5) around which cord is folded, 5 mm. thick. Made of light wood. b, Mangaian stick needle (Bishop Mus., C2807): length, 235 mm.; middle thickness. 6 mm.; length of cord loops (1), 500 mm.; clove-hitch attachments (2, 2). c, European netting needle with closed end (1) and middle prong (2). d, gauge (Bishop Mus., C2807): length, 178 mm.; width, 12 mm.; thickness, 3 mm. e, gauge (Bishop Mus., C2808): length, 156 mm.; width, 25 mm.; thickness, 5 mm.

Figure 141.—Netting needles and gauges. a, Tahitian netting needle collected by G. Bennet before 1829 (British Mus., 7052): length, 417 mm.; middle width, 40 mm.; length of opening (1, 2), 44 and 52 mm.; length of slits (3, 4), 40 and 31 mm.; thickness of side edges, 26 mm.; middle thin portion (5) around which cord is folded, 5 mm. thick. Made of light wood. b, Mangaian stick needle (Bishop Mus., C2807): length, 235 mm.; middle thickness. 6 mm.; length of cord loops (1), 500 mm.; clove-hitch attachments (2, 2). c, European netting needle with closed end (1) and middle prong (2). d, gauge (Bishop Mus., C2807): length, 178 mm.; width, 12 mm.; thickness, 3 mm. e, gauge (Bishop Mus., C2808): length, 156 mm.; width, 25 mm.; thickness, 5 mm.

page 224 (77, p. 295) and in Hawaii for mending nets with a fine mesh (63, pp. 107, 108). Mesh gauges were termed kauta, except in Mangaia where the local name was una. They are flat pieces of wood, usually ironwood, with sharpened edges and are made in different sizes to fit the meshes of various kinds of net (fig. 141, d, e).
The mesh was termed mata except in Mangaia, where it was termed koata. The length of a net was not counted in linear terms, but in the number of meshes. The meshes were counted by the knots in a row and thus included the intermediate meshes above or below, so that each additional knot counted as two meshes. Thus five knots were ten meshes, and an additional five knots counted as 20 meshes which were termed a mata. Hence in making a particular net, the craftsman counted the knots and stopped his row when he
Figure 142.—Netting technique, Mangaia. a, 1, 1, upper cord for making upper part of first row of meshes, held by assistant; 2, 2, lower cord for completing meshes, attached to needle (2′); upper cord (1) on left is looped over suspensory cord (3) and is drawn down into a loop against upper edge of net gauge (4); lower cord (2) on left passes behind gauge and up on front to meet upper loop when meshing knot (5) is made; this is repeated until first row is completed. Meshing knot is shown on right where netting needle carrying cord (2) passed through loop (6) formed by upper cord, made a loop to right (7) in passing over both limbs of upper loop, passed back under them and through its own loop; left thumb in front and left forefinger at back hold first crossing against upper edge of gauge, and after needle passes through its own loop knot is drawn taut against gauge. b, reversed upper row (1) showing backs of netting knots with last mesh (5) now on left with cut off end (1′) of upper cord and lower cord (2) descending from last knot. Gauge (4) is placed with its upper edge against lower end of completed first row meshes, lower cord (2) is passed behind gauge (4), and brought up in front to engage first lower loop (5) in orthodox netting knot (6) which again is shown in detail on right of figure.

Figure 142.—Netting technique, Mangaia. a, 1, 1, upper cord for making upper part of first row of meshes, held by assistant; 2, 2, lower cord for completing meshes, attached to needle (2′); upper cord (1) on left is looped over suspensory cord (3) and is drawn down into a loop against upper edge of net gauge (4); lower cord (2) on left passes behind gauge and up on front to meet upper loop when meshing knot (5) is made; this is repeated until first row is completed. Meshing knot is shown on right where netting needle carrying cord (2) passed through loop (6) formed by upper cord, made a loop to right (7) in passing over both limbs of upper loop, passed back under them and through its own loop; left thumb in front and left forefinger at back hold first crossing against upper edge of gauge, and after needle passes through its own loop knot is drawn taut against gauge. b, reversed upper row (1) showing backs of netting knots with last mesh (5) now on left with cut off end (1′) of upper cord and lower cord (2) descending from last knot. Gauge (4) is placed with its upper edge against lower end of completed first row meshes, lower cord (2) is passed behind gauge (4), and brought up in front to engage first lower loop (5) in orthodox netting knot (6) which again is shown in detail on right of figure.

page 225 reached the number required. Of a kukuti net it is said, "E ono mata ka oti te kukuti" (With six mata [120 meshes] a kukuti net can be finished). In describing nets, however, I shall follow the European linear measure.

Netting with the shuttle and mesh gauge is termed ta kauta to distinguish it from the Aitutaki technique in which a large ball (pokai), used without a shuttle, is termed ta pokai. Netting with the shuttle is similar to the Mangaian technique with the stick needle, and the same mesh knot is used throughout Polynesia.

To commence, the number of loops attached to the stick needle are such as will permit the thickness of the needle with the accompanying loops to pass readily through the required size of mesh. Before attaching the first loop, a length of twine that will be sufficiently long to form the upper half of the meshes of the first row is left free. The first row of meshes is formed over a cord loop that is stretched taut between the big toes. The free end of the twine is held by an assistant who keeps it taut as the meshing knots are made against the upper edge of the mesh gauge. The technique of making the first row of meshes is shown in figure 142, a.

When the first row that forms the depth of the net is reached, the assistant is dispensed with and any extra length of the free cord that has not been used is cut off. The supporting cord between the toes is reversed so that the last mesh which finished on the right is now to the left. A second row is started from the left (fig. 142, b) and so by reversing the suspensory cord on the completion of each row, the netting is continued by rows until the full length of the net is completed.

I have described the Aitutaki method (ta pokai) with technical details (70, pp. 281-285), but the main principles are repeated here. One end of a long length of cord is wound into a large ball; the other end is wound into a smaller ball, which contains sufficient cord to form the upper part of the loops of the first row, as does the free end of cord in the Mangaian method. A suspensory loop with another length of cord is stretched between the big toes, or a single length may be tied to each toe. Referring to the cord from the smaller ball which is held by an assistant as the upper cord and that from the large ball as the lower cord, the junction between them is arbitrarily defined by the upper edge of the mesh gauge when it is placed against the cord to form the first mesh. The upper cord is passed over the suspensory cord and the lower cord under the mesh gauge. Details are shown in figure 143.

When the first row is finished, the supporting cord is reversed (as in the Mangaian method) and additional rows are made in the same manner, the only difference being the form of netting knot. Other details already described are attaching a new cord (70, p. 285) and closing the ends and bottom of a scoop net (70, p. 286).

page 226

The Aitutaki method is novel in that the large ball of twine used cannot pass through the meshes. However the ball is passed through large loops outside the meshes. Each of the two large loops makes it possible to form two half-hitches on the lower loop of the mesh above, and the completed knot forms a reef knot. The technique seems peculiar to Aitutaki, and the Aitutakians are rather proud of the fact that their method differs from that of the other Cook Islands. Whether the technique is a local invention or is diffusion from some other area, I do not know.

Figure 143.—Netting technique, Aitutaki. a, 1, 1, upper cord with small ball; 2, 2, lower cord with large coil; 3, suspensory cord; 4, mesh gauge; upper cord (1) is looped over suspensory cord (3) and a loop (5) brought down to upper edge of gauge (4); lower cord (2) passes behind gauge and passes up in front where a loop (6) is passed over upper loop (5) without attempting to pass the large ball through upper loop (5). b, loop (6) is drawn out until it is large enough for the large ball to pass through it; right hand is passed through from below when hand twists left side of loop over to right. c, loop (6) is drawn taut against upper edge of gauge. d, another loop (7) is formed with lower cord (2) and passed under loop (5) of upper cord. e, loop (7) is drawn over upper loop (5) and enlarged to admit passage of large ball; it will be noted that first lower loop technique (6) has now resolved into a half-hitch around upper loop (5). f, large ball with cord (2) is drawn up through open loop (7) to make a second half-hitch. g, second loop is drawn taut to form second half-hitch (7) and in this stage, transverse loop formed by lower cord is shown below upper edge of gauge. h, when cord is drawn more taut, transverse loop below edge of gauge is forced upward (8) and knot resolves itself into a reef knot, which is shown open in first mesh (8) and closed in second (9).

Figure 143.—Netting technique, Aitutaki. a, 1, 1, upper cord with small ball; 2, 2, lower cord with large coil; 3, suspensory cord; 4, mesh gauge; upper cord (1) is looped over suspensory cord (3) and a loop (5) brought down to upper edge of gauge (4); lower cord (2) passes behind gauge and passes up in front where a loop (6) is passed over upper loop (5) without attempting to pass the large ball through upper loop (5). b, loop (6) is drawn out until it is large enough for the large ball to pass through it; right hand is passed through from below when hand twists left side of loop over to right. c, loop (6) is drawn taut against upper edge of gauge. d, another loop (7) is formed with lower cord (2) and passed under loop (5) of upper cord. e, loop (7) is drawn over upper loop (5) and enlarged to admit passage of large ball; it will be noted that first lower loop technique (6) has now resolved into a half-hitch around upper loop (5). f, large ball with cord (2) is drawn up through open loop (7) to make a second half-hitch. g, second loop is drawn taut to form second half-hitch (7) and in this stage, transverse loop formed by lower cord is shown below upper edge of gauge. h, when cord is drawn more taut, transverse loop below edge of gauge is forced upward (8) and knot resolves itself into a reef knot, which is shown open in first mesh (8) and closed in second (9).

Long Nets

Long nets are termed a'oroa (long line) in Aitutaki and tu kanae in Rarotonga, where they are used for catching mullet (kanae). Each net has an upper and lower rope threaded through the marginal meshes. Savage (57) states that the upper part of the net is termed karii-i-runga (upper kari'i) and the lower part, karii-i-raro (lower kari'i), but it is probable that these terms refer to the two ropes. The upper rope carries the floats (pouto or uto for short) page 227which are short cylindrical lengths of hibiscus or po'utukava from which the bark is removed. Holes are bored through the middle longitudinally to admit a cord. The po'utukava has a pith canal and the holes are easily made. In Aitutaki, the float cord is tied to the upper rope and a float threaded; then two clove-hitches at short intervals are made around the rope; then another float with clove-hitches; and so in sequence for the length of the upper rope (70, p. 295, fig. 256).

Sinkers (kari'i) are stones wrapped in coconut leaf stipule (kaka) and tied at intervals to the bottom rope. The wrapping projects beyond the stone at each end, leaving sufficient length for these ends to be lashed to the rope; and a third lashing passes around the middle of the package (70, p. 295, fig. 257). No poles were used at the two ends.

In Aitutaki and Rarotonga, long nets are used in the lagoon by a number of people in canoes, two or more canoes looking after the net. One end is held by one canoe, and the canoe loaded with the net pays it out carefully as it paddles in a curve across the channel. The occupants of other canoes start beating the water with their paddles or poles at some distance and gradually work toward the net, driving fish before them. The canoes spread out in a semi-circle to prevent the fish escaping at the sides. When the fish reach the net or just before they reach it, the canoes with the net ends quickly paddle toward each other to complete a circular enclosure. Other canoes dash up to various parts of the net, following it to lift the bottom rope over rocks or to block channels that might allow the fish to escape. As the size of the enclosure decreases, men jump out of their canoes and over the net rope to catch the fish with spears and scoop nets. There is great rivalry for the largest catch, but this is merely sport. After the fish are procured, the net is carefully drawn in and folded in the carrying canoe, and another site is selected for setting the net. In this manner, the length of the lagoon is fished.

In Aitutaki, there is a large lagoon and the long net is much used during seasons when large shoals of fish enter the lagoon. In Rarotonga, there is a good stretch of suitable lagoon between the villages of Ngatangiia and Titikaveka. I imagine that the other islands are not suited to this form of net. The principle is to create a barrier toward which shoals of fish are driven and then enclosed to prevent escape. The process of drawing the net onto a beach is not used as the lagoon bottom is unsuitable. A scout is sent ahead in his canoe to locate schools of fish and the tactics followed depend upon knowledge of the suitable places to set the net.

A shorter net in Rarotonga was termed a rerekue. These were evidently well made and considered valuable for those belonging to chiefs were given proper names. Thus Rua's net was named Tua-kere-maro, Taupine's net Te 'Uti-marumaru, and the net of Ara-pa'i was named Tukia.

page 228
Short Nets

Short nets with an upright pole at each end were much in vogue and were easily managed by two men while others drove fish into them. When fish struck the net, the two men brought the two poles quickly together and so enclosed the fish though many were enmeshed. The different islands had different names for the same type of net and distinctive names were applied according to length and size of mesh.

The best known were of the type termed nariki in Mangaia and Rarotonga. Gill (30, p. 9) writes that the Mangaian nets were invariably six yards long and four yards deep and that the meshes were so small that only the very tips of the fingers could be admitted (fig. 144). In Mangaia, the method of fishing with them was termed nariki, after the name of the nets. The two bearers of the net set them near the shore to catch small fish that could not escape through the fine meshes. Thirty to 40 men formed an oval to drive the fish and they were termed rau tangata. Rau (leaf sweep) refers to the men (tangata) who took the place of a leaf sweep. If fewer men were employed, they filled in the wider spaces in the drive with a leaf sweep (rau rakau). The method was used in the daytime; and as the men saw likely looking rock hideouts, they groped for refugee fish. In Mangaia, this technique, if used at night, was termed rau tata, but hand groping was not done.

Figure 144.—Nariki net, Mangaia: 1, 1, end poles, each held by one man; 2, top cord, threaded through marginal meshes; 3, 3, side cords, also threaded through marginal meshes and tied to poles by separate lashings at intervals. Net is much deeper than distance from top cord to ground, hence shows in folds.

Figure 144.—Nariki net, Mangaia: 1, 1, end poles, each held by one man; 2, top cord, threaded through marginal meshes; 3, 3, side cords, also threaded through marginal meshes and tied to poles by separate lashings at intervals. Net is much deeper than distance from top cord to ground, hence shows in folds.

Nets with a larger mesh than the nariki had were termed tata in Mangaia. Gill (30, p. 10) says that these nets were coarse, six yards long, and used daily to catch larger fish. Fishing with this net was termed ko tata. It was used in the deeper channels toward the reef. The two net holders set it across the channel and narrowed (ta pa'ipa'i) it by overlapping the slack and tying ('itikitiki) to suit the width of channels narrower than the net. Five or six men beat the water to drive the fish down the channel. Two divers (ta'aro), stationed near the net on either side of the channel, kept watch; and when they saw a school of fish pass by, they dived in behind it to drive it into the net. The fish were driven in from the shore side in the daytime at low tide or when the tide was beginning to come in, and, as the tide deepened, the net worked nearer to the shore.

page 229

In Rarotonga, nets about 10 yards long, termed toto, are used on the outer lagoon and are set parallel with the reef. Men with sticks drive fish into the net from the shore and from the sea side of the net.

Nets shorter than the nariki are also used. In Aitutaki, they are about eight feet long and six to seven feet deep and are termed tutu-rua (two poles) after the pole at each end. This net has quite a bulge in the middle. Two men carry the net and set it in a suitable place when a school of fish is seen, and others drive the fish into the net. Lengths of leaf sweep may be stretched out from the net on either side to direct fish into the net. Shoal fish such as uoa are caught near the white sandy beaches (kena) of the smaller islets on the reef. Other fish caught are vete, ka'a, and kanae (mullet), usually at full tide. In the medium tides, rakoa and paoa are caught. In February, after thunder at night, large schools of maito are caught near the beaches (70, p. 293, fig. 255).

In Rarotonga, the tana'o net, about five yards long, is used near the shore to catch fish in shoals (na'o).

Gill (30, p. 10), after describing the nariki and tata nets of Mangaia, states that a smaller net, only two yards long, was termed a kukuti. The same name was applied to a small scoop net.

A small, fine-meshed net with handles, which was used in Rarotonga, was termed a ta'iti. It is said to have had floats and sinkers and was probably a small form of the nariki net after the style of the Samoan shrimp net (73, pp. 479, 480, pl. XLVI, A).

Scoop Nets

A scoop net is in the form of a bag that is attached to an elliptical hoop frame with or without a handle.

Scoop Nets without Handles

Large nets with a frame six to eight feet long and from 2.5 to 3 feet wide across the widest part of the frame are used throughout the island under various names (fig. 145). In Aitutaki and Atiu they are termed 'opai; in Mauke, kupenga toto; in Rarotonga, pokipoki; and in Mangaia, pdei'ei. The frame (tutu) is formed of two rods lashed at both ends and spread out into the form of an ellipse by a crossbar.

The net is held by the crossbar with the fisherman's elbow braced against the angle formed by the butt ends of the frame elements. It is used in weirs of the V-shaped type to block the entrance passage after fish enter the enclosure and is also used to scoop up the imprisoned fish within the enclosure. Two Aitutakian methods of use have been described (70, pp. 291, 292) as follows.

The eki method, which is used on the reef on moonlight nights at low tide, when fish are believed to be asleep in the channels, is used by one man. He places his net in position at a suitable place in the channel near some rocks; and, with a pole in his left hand, he page 230prods (uruuru) among the rocks to startle the fish into his net. He holds the slack bottom of the net with his right hand against the crossbar of the frame. When a fish darts against the taut net the impact is felt, and the slack is released and the net brought up. The fish caught are kauva'a, u'u, nanue, kovia, and sometimes urua.

The akaoro ngaika u'u (driving schools of parrot fish) method requires two or more assistants to drive the schools of parrot fish (u'u) or other fish down the channel into the net placed in a suitable place in the channel. This method is used in the daytime when the tide is coming in (tai maene) or when it is going out (tai tuku), because at full tide the channels are too deep.

Figure 145.—Scoop net without handle ('opai), Aitutaki. a, length, 7 feet 9 inches; greatest width, 2 feet 8 inches; length of crossbar (pukei), 15 inches; distance of crossbar from proximal end, 14 inches; depth of bag net, 5 feet 8 inches; frame (1) of two branches of ironwood each 9 feet 5 inches long, about 1 inch thick at near end and 0.5 inch thick at far end. b, attachment of net to frame (1) by one cord (2) which passes through marginal meshes and at intervals is fastened to frame and cross bar by clove-hitches (3). c, net attachment by two cords, one cord (2) passing through marginal meshes and second cord (3) making loose spiral turns around frame (1) and marginal cord (2).

Figure 145.—Scoop net without handle ('opai), Aitutaki. a, length, 7 feet 9 inches; greatest width, 2 feet 8 inches; length of crossbar (pukei), 15 inches; distance of crossbar from proximal end, 14 inches; depth of bag net, 5 feet 8 inches; frame (1) of two branches of ironwood each 9 feet 5 inches long, about 1 inch thick at near end and 0.5 inch thick at far end. b, attachment of net to frame (1) by one cord (2) which passes through marginal meshes and at intervals is fastened to frame and cross bar by clove-hitches (3). c, net attachment by two cords, one cord (2) passing through marginal meshes and second cord (3) making loose spiral turns around frame (1) and marginal cord (2).

The Mangaian pa'ei'ei net is from 5 to 6 feet long and two methods were described by informants as follows.

The pa'ei'ei method takes its name from the net and is the same as the second Aitu-takian method. The holder of the net has a stick which he moves about in any space between the net and the side of the channel preventing fish from escaping through the gap.

The ta'ata'a method is used on the reef to catch fish swimming about in the waves. When fish are feeding on the reef, they come in on the waves and return with the back wash. The fisherman places his net behind the wave so that the fish are swept into the net on the back wash. He uses his feet and a stick to direct fish into the net. As a wave breaks, he stands on one foot to lessen resistance to its force.

The Rarotongan pokipoki net was similar to the Aitutakian 'opai, but I saw one attached to the frame by a continuous chain knot instead of spaced clove-hitches. This net was 6 feet 2 inches long and 3 feet 3 inches across at its widest part, and the crossbar was 2 feet 7 inches wide. The stick used to guide fish is termed a koko.

In Mangaia, a small scoop net about 3 feet long was termed a kukuti. In Mangaian terminology, the length was said to be a tapa'e, the distance from page 231the outstretched middle finger at arm's length to the middle of the chest. The kukuti was used in narrow channels that suited its size. The method of fishing with it was termed pato'ito'i, which differs in no way from the method of using the larger pa'ei'ei net.

Scoop Nets with Handles

The best known scoop net with a handle is the flying-fish net, generally termed 'uata. It is an oval frame (tutu) formed of two ironwood rods attached to a long handle (kakau) by their thicker ends and tied together at the other ends to form a distal point (katatai). A wooden crossbar (pukei, Aitutaki) acted as a spreader to maintain the oval form of the frame. In Atiu and Mauke, the crossbar is termed tiare after the wood used, and this name has evidently displaced an older specific term. A fairly deep bag net is attached to the frame and the crossbar. In Aitutaki, the attachment is by means of a circumferential cord passing through the marginal meshes and taking single half-hitches around the frame at short intervals. In Atiu, a circumferential cord is confined to the marginal meshes and another cord passes around the circumferential cord and the frame in continuous spiral turns.

Figure 146.—Flying-fish net ('uata), Atiu. a, frame and net: 1, handle (kakau); 2, frame (tutu) of two branches of ironwood lashed to handle, diverged by spreader (3), and distal ends lashed together; length of frame opening between spreader and distal end, 41 inches; greatest width, 21 inches; 3, spreader or crossbar (tiare) lashed to handle end and sides of frame with neat fly-flap pattern in sennit (fig. 147). 4, sennit lashing to strengthen spreader. b, lashing of frame rods to handle: 1, handle; 2, frame rods, trimmed flat on one side to rest against handle and lashed with three sets of transverse turns which finish with a cross turn (3) passing around lashing between handle and branch on each side. c, sennit lashing to strengthen spreader: two complete transverse turns (1) are made around side branches and over handle, and parts between wooden elements are seized (2).

Figure 146.—Flying-fish net ('uata), Atiu. a, frame and net: 1, handle (kakau); 2, frame (tutu) of two branches of ironwood lashed to handle, diverged by spreader (3), and distal ends lashed together; length of frame opening between spreader and distal end, 41 inches; greatest width, 21 inches; 3, spreader or crossbar (tiare) lashed to handle end and sides of frame with neat fly-flap pattern in sennit (fig. 147). 4, sennit lashing to strengthen spreader. b, lashing of frame rods to handle: 1, handle; 2, frame rods, trimmed flat on one side to rest against handle and lashed with three sets of transverse turns which finish with a cross turn (3) passing around lashing between handle and branch on each side. c, sennit lashing to strengthen spreader: two complete transverse turns (1) are made around side branches and over handle, and parts between wooden elements are seized (2).

I have described specimens from Aitutaki and Mauke (70, pp. 287-289, figs. 247-249), but a fine specimen from Atiu is here figured in plate 10, C and figure 146. The Atiu net is neatly lashed with the fly-flap pattern (fig. 147) at the junctions of the crossbar with the frame and handle.

The 'uata net is used at night for catching flying fish (maroro) in the open sea. The method is termed rama maroro (to torch for flying fish) after the page 232coconut-leaf torches (rama) which are used. A number of torches are carried in the canoes, a fresh one being lighted from the one that is burning out. For the usual small fishing canoe, a crew of three men is required, one to wield the net, one to hold the torch, and one to sit in the stern and paddle the canoe. The flying fish are attracted by the light, and as they fly through the air near the bow of the canoe the net wielder dexterously intercepts them and empties the mouth of the net into the canoe. The manual action of sweeping with the net is termed ta'ei and painga and the net wielder who stands in the bow is called "the man who netted flying fish" (tangata ta'ei maroro). The act of using a lighted torch is tu rama, and the torch holder who stands behind the net wielder is called "the man holding the torch" (tangata mou rama) or "holder of the fire" (tutu a'i).

Figure 147.—Fly-flap lashing technique on Atiu net. a, end of handle with crossbar placed against end thus dividing bar into right and left half; one end of cord is fastened to end of handle with a running noose; cord passes obliquely upward from left and takes a complete turn (1) around right bar, crossing on top of bar. b, cord descends obliquely on back to left of handle and loops around front to right. c, from right side of handle cord ascends obliquely at back to form complete turn (2) around left bar crossing on top and descends obliquely to right. d, cord loops around back of handle to left, ascends obliquely to right below previous oblique turn, and commences a second (3) turn around right arm. e, cord completes turn (3) around right arm, descends obliquely on back to left, and loops (4) around front of handle. f, from right of handle, cord passes obliquely upward at back toward left to make second complete turn (5) around left bar, and then passes obliquely downward to right (6) below previous oblique turn. g, sequence of turns is now established; complete turns alternating around the crossbar limbs and half turns around handle, alternately on front and back. Crossings are kept in line on top of bar and sides of handle. Sequence results in complete pattern (g) which is fixed by some transverse turns (7) which cover cord end.

Figure 147.—Fly-flap lashing technique on Atiu net. a, end of handle with crossbar placed against end thus dividing bar into right and left half; one end of cord is fastened to end of handle with a running noose; cord passes obliquely upward from left and takes a complete turn (1) around right bar, crossing on top of bar. b, cord descends obliquely on back to left of handle and loops around front to right. c, from right side of handle cord ascends obliquely at back to form complete turn (2) around left bar crossing on top and descends obliquely to right. d, cord loops around back of handle to left, ascends obliquely to right below previous oblique turn, and commences a second (3) turn around right arm. e, cord completes turn (3) around right arm, descends obliquely on back to left, and loops (4) around front of handle. f, from right of handle, cord passes obliquely upward at back toward left to make second complete turn (5) around left bar, and then passes obliquely downward to right (6) below previous oblique turn. g, sequence of turns is now established; complete turns alternating around the crossbar limbs and half turns around handle, alternately on front and back. Crossings are kept in line on top of bar and sides of handle. Sequence results in complete pattern (g) which is fixed by some transverse turns (7) which cover cord end.

In Mangaia, the hoop frame of the flying-fish net is inclined forward at an angle with the handle. This variant technique is shown in figure 148, a, b.

Torching is usually carried out by a number of canoes on dark nights. In Mangaia, I was told that a number of men I saw gathered on the beach in the early evening were going out to torch for flying fish. As no move was made, I asked when they were going. A man pointed to the new moon a little above page 233the western horizon and said, "When that sinks we will go out, for it is no use lighting torches when the moon is up." Torching is organized so that each canoe receives a fair chance. They are assembled outside the reef, opposite the reef channel through which they have passed, and when they reach the grounds they deploy into line. The leading canoe has the best chance for fish, but when it has burnt out one torch, it drops to the rear and the second canoe leads the procession for the duration of one torch. The sequence continues until the original first canoe again leads the van; and the sequence is repeated.

Figure 148.—Mangaian flying-fish net. a, back of frame: 1, handle, length 10 feet 2 inches; diameter, 1.5 inches; 2, crossbar lashed to front end of handle, length, 17 inches, diameter, 0.75 inch; 3, 3, ironwood branches, 3 feet 6 inches long and 0.6 to 0.4 inch in. diameter, thick ends crossed on handle 6 inches from end and lashed to it, then passed over front of crossbar and lashed to its ends, bar serving as spreader. Distal ends were brought together and tied, frame from butt ends to tips being 2 feet 6 inches long and the greatest width, 1 foot 8.5 inches. b, side view of frame: hoop branches (3) tied to back of handle (1) and in front of crossbar (2) gives hoop a forward angle to axis of handle. A bag net (4), 19 inches deep and of same length at bottom as at opening is attached to sides of hoop and crossbar. c, Mangaian kukuti hand net: two ironwood branches with smaller branches jutting out at appropriate angles were selected. Larger branches (1, 1) were cut short beyond branch junction on acute angle side and at 4 inches on other side to provide a handle, the two being trimmed flat on opposing surfaces and lashed together. Smaller branches (2, 2) were 0.5 inch thick at junction and 0.2 inch at their ends, their total length 3 feet. When ends were drawn together and tied, length of hoop so formed was 2 feet 8 inches and greatest width 15 inches. A bag net, 3 feet 4 inches deep and 18 inches wide at bottom was attached to frame by two cords, one passing through marginal meshes of net opening and a second passing around hoop and marginal cord in loose spiral turns.

Figure 148.—Mangaian flying-fish net. a, back of frame: 1, handle, length 10 feet 2 inches; diameter, 1.5 inches; 2, crossbar lashed to front end of handle, length, 17 inches, diameter, 0.75 inch; 3, 3, ironwood branches, 3 feet 6 inches long and 0.6 to 0.4 inch in. diameter, thick ends crossed on handle 6 inches from end and lashed to it, then passed over front of crossbar and lashed to its ends, bar serving as spreader. Distal ends were brought together and tied, frame from butt ends to tips being 2 feet 6 inches long and the greatest width, 1 foot 8.5 inches. b, side view of frame: hoop branches (3) tied to back of handle (1) and in front of crossbar (2) gives hoop a forward angle to axis of handle. A bag net (4), 19 inches deep and of same length at bottom as at opening is attached to sides of hoop and crossbar. c, Mangaian kukuti hand net: two ironwood branches with smaller branches jutting out at appropriate angles were selected. Larger branches (1, 1) were cut short beyond branch junction on acute angle side and at 4 inches on other side to provide a handle, the two being trimmed flat on opposing surfaces and lashed together. Smaller branches (2, 2) were 0.5 inch thick at junction and 0.2 inch at their ends, their total length 3 feet. When ends were drawn together and tied, length of hoop so formed was 2 feet 8 inches and greatest width 15 inches. A bag net, 3 feet 4 inches deep and 18 inches wide at bottom was attached to frame by two cords, one passing through marginal meshes of net opening and a second passing around hoop and marginal cord in loose spiral turns.

In Atiu and Mauke, double canoes were formerly used for torching flying fish. The net wielder stood with a foot on the bow of each canoe. With four-oars pulling, these canoes must have made good speed and traveled longer distances to sea than did single canoes.

page 234

Flying fish are sometimes caught in the lagoon, and the flying-fish net may be used in torching to scoop other fish out of the water.

A net with a shorter handle, used in Aitutaki, and termed a ngake (70, pp. 292, 293, pl. 253) is a smaller edition of the flying-fish net, but the bag net is attached to the frame by a circumferential cord through the marginal meshes and a spiral cord around the first cord and the frame as in the Atiu flying-fish net. It is used at night with torches to catch small fish such as i'e (piper) and muro and shoal fish such as koamo and tikoami. Fish may be driven toward the net with coconut leaf mauru and the net may also be used as a scoop without a handle for fish enclosed by leaf sweeps.

The Mangaian kukuti net is ingeniously made from two branched branches and does not require a crossbar spreader (fig. 148, c).

A net frame in the London Missionary Society's collection in the British Museum, attributed to the Hervey Islands, is peculiar in that the handle curves down below the hoop frame to be attached to its distal end and there is no crossbar (fig. 149).

Figure 149.—Scoop net frame (British Mus., 514). a, bottom view: handle (1) of dark wood 1,040 mm. long including curve; diameter, 29 mm.; frame (2) is one piece of wood bent in an oval with two ends lashed to underside of handle (3) with sennit 3 mm. wide; wide bend of hoop attached to end of handle (4) with sennit 5 mm. wide; hoop length, 420 mm.; hoop greatest width, 320 mm. b, side view, showing curve of handle: end of handle (4) notched for hoop lashing, c, enlarged view of end lashing (4).

Figure 149.—Scoop net frame (British Mus., 514). a, bottom view: handle (1) of dark wood 1,040 mm. long including curve; diameter, 29 mm.; frame (2) is one piece of wood bent in an oval with two ends lashed to underside of handle (3) with sennit 3 mm. wide; wide bend of hoop attached to end of handle (4) with sennit 5 mm. wide; hoop length, 420 mm.; hoop greatest width, 320 mm. b, side view, showing curve of handle: end of handle (4) notched for hoop lashing, c, enlarged view of end lashing (4).

Driving Fish

Fish are driven into nets by various methods. Beating upon the water with poles is termed papa and thrusting with the pole into pools and crevices is koko. In Aitutaki, a five foot length of coconut leaf termed a mauru is used to beat the water. Leaflets are stripped off the thicker end to form a handle and the leaflets at the other end are split on either side of the leaflet midribs page 235which are then pinched off leaving the soft parts of leaflets attached to the leaf midrib. Stones picked up from the lagoon bottom are thrown behind the fish to keep them moving toward the net. Noises are made with the hands. Beating upon the thighs is termed pokipoki and beating the water with two cupped hands brought together is porutu.

Angling

Fishing with hook (matau) and line (a'o) is still a favorite occupation of the men. Women sometimes say that the men go fishing to avoid the chores around the home or in the plantations. Certain it is that when the weather is not rough a fleet of canoes can always be seen in the daytime fishing on some selected ground off the coast. Wherever the fishermen's local seasonal knowledge leads them, the cluster of specks seems to remain stationary all day.

The lines are made like the cord for nets (p. 223) and some are long for deep-sea fishing. The simplest form of implement attached to a line for securing fish is a gorge.

Gorges

Gorges (tara) are made from a short length of black ngangie wood sharpened at both ends and with a line tied to the middle (fig. 150, a). The gorge is held in the same long axis as the line and inserted into the bait. When cessation of movement of the line indicates that the bait has been swallowed, the line is pulled and the drag on the middle of the gorge causes it to turn cross-ways in the fish's stomach (kua tarava te tara i roto i te kopu). The fish is thus effectively hooked and can be drawn up. In Mangaia, the gorges are straight or curved.

The straight gorge (tara pu'i) is used for catching sea eels (pu'i) in holes in the rocks within the lagoon. The bait is the small panoko fish. After the bait is tied on, one point of the gorge is stuck into a length of cane (kaka'o), which makes it possible for the bait to be thrust into likely looking holes in the rocks. When the fisherman feels the bait taken, he draws away the cane but leaves the line slack until cessation of movement indicates that the bait has been swallowed. Sea eels are voracious and swallow the bait without playing with it. This form of fishing is termed mangai pu'i. In Rarotonga, the gorge is called te matau o te pu'i (the hook of the sea eel).

The curved gorge (tara karore) receives its name because it is used to catch the karore, a small fish frequenting the outer side of the reef. The karore pulls and shakes the bait before swallowing it, so the curved gorge is believed to be better than the straight gorge for catching this fish. Two gorges are used, one attached with a two-inch length of line to each end of a spreader, consisting of two combined pieces of coconut leaflet midribs (kikau) 10 inches long. The line is tied to the middle of the stretcher and a short length extends page 236below to which a stone sinker is attached with a piece of hibiscus bark (kiri'au). Each gorge has a short length of thin line attached to it for tying on the bait, ungakoa or hermit crab (ungaunga).

The line is lowered until the sinker touches the bottom, but the bait is suspended above the bottom. When the karore fish pulls on the bait, which is plainly seen in the clear water, the line is slightly jerked. The fisherman waits until a fish takes the other bait before drawing in the line.

Other fish caught with a gorge are takataka and rari. If octopus bait could not be procured by deep sea fishermen, they fished for karore from their canoes to obtain a substitute.

Figure 150.—Gorge and toko hook. a, Rarotongan straight gorge (Bishop Mus., C2845): made of hard wood (1), 80 mm. long, round in section, 6 mm. thick in middle, and pointed at each end; thick two-ply bast cord (2) attached by doubling bast strands around it and plaiting into cord; another bast strip lashed around cord and gorge join; thin cord (3) usually attached to middle to tie bait on. b, toko hook, Mangaia (Bishop Mus., C2804): fork of hard ngangie wood; greatest length, 168 mm.; greatest outside width, 68 mm.; shank limb diameter below snood, 8 mm.; at bend, 14 mm.; slight shank knob (1) under lashing; point piece (2) curved and attached to point limb with transverse bast lashing. Modern, but made on old pattern.

Figure 150.—Gorge and toko hook. a, Rarotongan straight gorge (Bishop Mus., C2845): made of hard wood (1), 80 mm. long, round in section, 6 mm. thick in middle, and pointed at each end; thick two-ply bast cord (2) attached by doubling bast strands around it and plaiting into cord; another bast strip lashed around cord and gorge join; thin cord (3) usually attached to middle to tie bait on. b, toko hook, Mangaia (Bishop Mus., C2804): fork of hard ngangie wood; greatest length, 168 mm.; greatest outside width, 68 mm.; shank limb diameter below snood, 8 mm.; at bend, 14 mm.; slight shank knob (1) under lashing; point piece (2) curved and attached to point limb with transverse bast lashing. Modern, but made on old pattern.

Hooks

Hooks (matau) of the old types have gone out of use. They are rare even in Museums, and except for an occasional wooden shark hook, are not found in the islands.

The simplest form of hook, made from pandanus leaf thorns, is described by Gill (28, pp. 118, 119) as follows:

The serrated edges of a stout leaf are pared off; the narrow pieces are then carefully tied together with a bit of hibiscus bark, care being taken that there be at least two thorns or tiny fish-hooks on either side, and that these little hooks point upwards. The slit midrib of a long coconut frond furnishes the fishing rod—the thorny hooks being secured to the tapering end.

These hooks were used by children to catch gobies (kokopu) in the fresh water streams and pools, with pieces of fresh water shrimp as bait. The mode of angling is said to have been invented by the Mangaian ancestor, Tarauri, who also earned fame by conquering the seven dwarf sons of Pinga in a dart throwing match. In recent years, children make hooks out of thorns of the introduced lime which are tied at an angle to a length of coconut leaflet midrib.

page 237

In Mangaia, small and medium-sized hooks were made of coconut shell (upu) or the shell of a Turbo (ariri) which Gill (30, p. 212) states was Turbo petholatus. The shells were cut on their natural curve, which supplied a curve to the hooks. A chip of chert (ruarangi) was used to cut the shell and Gill (30, p. 65) states that men always carried a chip of this material in their loin cloth to serve as a knife. In islands where chert was not present, a chip of basalt was probably used.

Bligh (46, p. 134) records that he obtained fishhooks made of turtle shell in Aitutaki, but none of these have come to light and there is no information as regards their exact shape. Beasley (4, plate LIII, A) figures four pearl-shell hooks attributed to the Cook Islands with snoods made of pandanus fiber. As pearl shell, except by importation, was not present in the Cook Islands and as the natives never used pandanus fiber for snoods, the hooks figured must come from some atoll. From their shape, they appear definitely to belong to Pukapuka, from whence numbers of these hooks are sent to Rarotonga as curios.

Though the Mangaians no longer have coconut and marine shell hooks, informants knew quite a lot about the methods in which they were used. This is probably because the same methods apply to the use of trade hooks. The three methods described for within the lagoon and three for outside the reef are the following.

1.The ta ko'eko'e method receives its name from a bamboo rod (ko'e) which is used with coconut shell hooks in the channels in the reef. The fish caught are patuki, karau, and pa'oro.
2.Titomo (to enter or thrust) is the method used with a coconut-shell hook attached to a line from 2.5 to 3 inches long fastened to an ironwood rod about 4.5 feet long. The rod is thrust (titomo) into holes in the sides of the channels. The bait is hermit crab and ungakoa.
3.In the method called 'i koura ('i, to angle; koura, crayfish), a small coconut-shell hook on a line is used in deep holes where there are crayfish. Ground bait (paru) is cast into the hole and the baited hook lowered with it. If a fisherman gets two or three crayfish from one hole and receives no more bites, he passes to another likely hole.
4.

The 'i nanue ('i, angling for nanue) receives its name from the nanue fish. It is caught in daylight with hooks of coconut shell or ariri shell, outside the reef in a depth of 7 to 8 fathoms where the bait can be watched. Land and sea crabs are used for bait. Ground bait is used to attract fish to the hooks and is crushed crustaceans or pieces of fish. It is wrapped up in a leaf and placed on a flattish stone. The baited hook is laid on the ground bait and the line above the hook is wound around the stone, bait, and hook for a number of turns. The left thumb is laid transversely over the turns and the line reversed around the thumb. The reversed line takes one more turn around the stone and the loop formed after disengaging the thumb is twisted under and over the last turn in an overhand knot. This is sufficient to hold all together. It is dropped quietly over the side of the canoe. The stone is lowered to the required depth, and a sharp jerk frees the loop. The stone sinks to the bottom and the leaf opens out to free the ground bait with the baited hook in its midst. Ume are also caught by this method.

Nanue were caught off the reef with a bamboo rod and ground bait could be used.

5.The pato method used coconut shell or ariri shell hooks a little farther away from the reef in deeper water, where the bait could not be seen. The fish caught were atea, ka'uru, tavatava, paueue, and taratara'oka.page 238
6.The 'i ku (angling for ku) method is angling with the same small hooks, but in deep water—10 to 20 fathoms—on moonlight nights (po 'inamotea) at about the marangi night (16th) of the lunar month. The fish caught are ku-taraeo, ku-matanui, ponui, tavatava, tarao and tongare (tarangau, Rarotonga).
Wooden hooks

Wooden hooks were made in three forms: a medium-sized hook termed toko, the large so-called Ruvettus hooks, and large shark hooks.

The toko hooks of medium size were used for deep-sea fishing. I saw some in Rarotonga in the possession of the late Makea Ariki (Tinirau); but though the shank and point limbs were fairly old, the original wooden points had been replaced with modern steel fishhooks with the bends straightened to copy the old wooden points. They were made of a forked branch of ironwood which formed the shank and point limbs and the point consisted of an angular piece of hard wood, one part being lashed to the end of the point limbs so that the other pointed part projected inward toward the shank to form an unbarbed point. In Mangaia, the point was formed of hard black ngangie wood. It was stated that bone and boar tusks were also used in Mauke. A hook from Aitu-taki, which I figured (70, p. 306, fig. 267) as a toko, is 280 mm. long (11 inches) and has all the dimensions of a Ruvettus hook. Evidently some confusion has arisen between the two forms, but, to the best of my knowledge, the toko is a medium-sized hook for catching barracuda (manga) and other fish. It was similar to the Ruvettus hook but much smaller. A hook of this form, made for me in Managaia, is shown in figure 150, b.

The Ruvettus hook has a wide distribution, as shown by Gudger (37). It is usually associated with catching the castor oil fish (Ruvettus ruvettus; Cook Islands, vena), but apparently the manga was more important in the Cook Islands. Like the toko hook, the Ruvettus hook was made of a forked branch of ironwood and fitted with a wooden point in the same manner, but the Ruvettus was larger. It resembles the Ruvettus hooks of Rakahanga (75, pp. 164-170) and Tongareva (74, pp. 208, 209). The hook was used in deep sea fishing, and lines had to be very long to convey the hook to the bottom.

Fishing took place by day or night, and the depths fished ranged from 1,200 to 1,800 feet. A fairly heavy sinker was used, and, to save hauling it up, one of two ingenious ways was used to get rid of the sinker after it had served its immediate purpose. The sinker was attached to the hook with a strip of pandanus leaf which was fixed to the bend of the hook with a Blackwall hitch that automatically released when the stone touched bottom (fig. 151, a), or the hook was bound to the stone by a series of turns with a slip knot that was released by a jerk of the line at any required depth (fig. 151, c). The second method resembles that used for releasing ground bait in nanue fishing (p. 237).

A fisherman had to take out a load of stones and some pandanus leaves if he used the automatic release. An assistant termed utauta collected the stones page 239and loaded them into the canoe. He also helped to launch the canoe but did not accompany the fisherman. If the fisherman was lucky, he rewarded the utauta with a fish.

Figure 151.—Sinker release attachments. a, automatic release: free end (1) of pandanus leaf strip attached to sinker (b) is laid against lower part of hook (2) and strip passes over end, weight of sinker maintaining tension. When sinker touches bottom, tension is removed, and strip drops off. b, other end of pandanus strip, showing end knot (1) with strip split near knot; strip passes through split to form running noose which holds sinker (2); sinker is so fastened before upper end of strip is attached to hook as in a. c, jerk release: baited hook (1) is laid on the sinker (2), and some turns of line pass around hook and sinker; line is then looped around left thumb and makes a reverse turn around sinker; thumb is removed and loop so formed (3) is twisted over and under line in first part of a reef knot sufficient to hold weight of sinker.

Figure 151.—Sinker release attachments. a, automatic release: free end (1) of pandanus leaf strip attached to sinker (b) is laid against lower part of hook (2) and strip passes over end, weight of sinker maintaining tension. When sinker touches bottom, tension is removed, and strip drops off. b, other end of pandanus strip, showing end knot (1) with strip split near knot; strip passes through split to form running noose which holds sinker (2); sinker is so fastened before upper end of strip is attached to hook as in a. c, jerk release: baited hook (1) is laid on the sinker (2), and some turns of line pass around hook and sinker; line is then looped around left thumb and makes a reverse turn around sinker; thumb is removed and loop so formed (3) is twisted over and under line in first part of a reef knot sufficient to hold weight of sinker.

Day fishing and night fishing had distinguishing terms in Mangaia. Day fishing ('akareva 'oka) was carried on at a depth of four angarere (1,200 feet). The bait used was karore, and the fish caught were 'oka, 'apuku, ru'i-kai'ara, and angamea. Night fishing ('i manga, to angle for manga) was done 'in four to six angarere (1,200 to 1,800 feet). The bait was flying fish and the fish caught were manga, vena (Ruvettus), mango (shark), ru'i kai'ara, and manga kai'ara. The term kai'ara used with the last two fish is the Mangaian term for very large. The manga may be two finger spans (18 inches) in thickness and one and a half maro (9 feet) in length. Some manga are deep ('a'ano) and others more rounded (punupunu). The round bodied barracuda are termed manga tangi tuoro.

Shark hooks. A large wooden hook in the Oldman collection (54, vol. 48, pi. 37, no. 380) attributed to the Cook Islands and another in the British Museum attributed vaguely to Eastern Polynesia are figured by Beasley (4, pl. LIII) as coming from the Cook Islands. Confirmation of the general locality comes from Mr. Drury Lowe who collected two similar hooks on the island of Mauke. Mr. Lowe sent me outline drawings which conform in shape and size to those described by Beasley, the smaller one being just under 8 inches long and the larger 9.5 inches long. Lowe states that his hooks were made of ironwood root and were used for catching kura vena (Ruvettus). Though conveniently regarded as shark hooks, they were used for catching other large fish as well.

The hooks, which are well made from a dark wood, are in one piece with the point curved inward and have a well-formed shank knob. The very long page 240snood is composed of sennit which takes a couple of turns around the shank below the knob, and after turning upward, is secured with a lashing of fine sennit (3 mm. wide) in a well-executed lozenge design that follows the inaere pattern of Mangaia. Both the Oldman and British Museum hooks have this lashing pattern and it may be that both are from Mangaia. The long snoods are seized with fine sennit throughout their length and each ends in a loop which is also seized. Beasley (4, p. 37) says the Oldman hook was attached to the line without an intervening snood; but the presence of the loop at the end of the seized line shows that it was intended for tying to a longer line, hence the seized line in spite of its length must be regarded as the snood. Long snoods in two well-preserved hooks make a distinction, apart from shape and separate point, between these hooks and the Ruvettus hooks. Beasley (4, p. 37) states that the binding material of the Oldman hook was of twisted cord made of oronga bark, but Oldman (54, vol. 48, p. 24, no. 380) states that it was of "fine plaited sennet". My own notes on the British Museum hook show that the lashing and seizing material was fine sennit (braid of coconut-husk fiber). The British Museum hook has a stone sinker attached to the snood. The two hooks are shown in figure 152.

Figure 152.—Wooden shark hooks. a, length, 266 mm. (Oldman coll., 380). b, enlargement of snood lashing of a:1, shank knob; 2, thick sennit, part of snood, wrapped around shank below shank knob; 3, fine sennit lashing, in inaere pattern, to fix snood; 4, seizing, which continued to end of snood. c, length, 225 mm.; outside width below snood lashing, 143 mm.; thickness of shank limb below lashing, 28 mm., at bend, 45 mm., of point limb opposite lower end of snood lashing, 26 mm.: 1, shank knob; 2, snood wrapping; 3, fine sennit lashing similar to b; 4, seizing around snood; 5, sinker, in form of disc with diameter of 65 mm. and thickness of 45 mm.; circumference grooved for sennit which is lashed to snood; 6, eye formed at end of snood by seizing (British Mus., 1903-18).

Figure 152.—Wooden shark hooks. a, length, 266 mm. (Oldman coll., 380). b, enlargement of snood lashing of a:1, shank knob; 2, thick sennit, part of snood, wrapped around shank below shank knob; 3, fine sennit lashing, in inaere pattern, to fix snood; 4, seizing, which continued to end of snood. c, length, 225 mm.; outside width below snood lashing, 143 mm.; thickness of shank limb below lashing, 28 mm., at bend, 45 mm., of point limb opposite lower end of snood lashing, 26 mm.: 1, shank knob; 2, snood wrapping; 3, fine sennit lashing similar to b; 4, seizing around snood; 5, sinker, in form of disc with diameter of 65 mm. and thickness of 45 mm.; circumference grooved for sennit which is lashed to snood; 6, eye formed at end of snood by seizing (British Mus., 1903-18).

page 241
Incantations termed mauri were often recited to insure success in fishing for shark with the baited hook. A Rarotongan mauri appeals to Tangaroa.
Tangaroa e! Ta taua mango e! O Tangaroa! The shark of us two!
'Utia! Lift it up!
'Utia ki runga i te rangi. Lift it up to the heavens.
Titi ka ruerue. Ah, ah, it moves.
Tata ka raruerue. Oh, oh, it comes up.
Titi takere takere to! (rhythmic phrase.)
'Utia means to bring up on a line. The last line is a rhythmical phrase imitating the sound of the wooden gong and used here to denote success.

Beasley (4, p. 37, pl. LIII, B) figures a hook roughly made from hard wood that was found in a midden at Tupapa, Rarotonga. It has a rounded bend like a shark hook but no curve for the point. It may be an incomplete hook of the toko type to which a separate front piece had to be attached (fig. 153, a).

Figure 153.—Wooden hooks. a, roughly made with two shank knobs (1, 1) and straight point limb (2) without inward curve; length, 160 mm. (E. L. Gruning coll.). b, ornamental hook (Fuller coll. 330): 1, snood lashed around shank knob; 2, 2, barb-like projections on shank limb and point piece; 3, four cleatlike projections; 4, four rounded knobs in midline of outer side surface, with notched raised ridges (5) extending between them; 6, point piece with carved triangular projection at outer bend; 7, join between point piece and point limb, covered with fish skin and lashed with two sets of sennit in transverse turns. Beasley gives measurements as 22.5 inches long and greatest width, 13.5 inches.

Figure 153.—Wooden hooks. a, roughly made with two shank knobs (1, 1) and straight point limb (2) without inward curve; length, 160 mm. (E. L. Gruning coll.). b, ornamental hook (Fuller coll. 330): 1, snood lashed around shank knob; 2, 2, barb-like projections on shank limb and point piece; 3, four cleatlike projections; 4, four rounded knobs in midline of outer side surface, with notched raised ridges (5) extending between them; 6, point piece with carved triangular projection at outer bend; 7, join between point piece and point limb, covered with fish skin and lashed with two sets of sennit in transverse turns. Beasley gives measurements as 22.5 inches long and greatest width, 13.5 inches.

Ornamental hooks

A very large hook of the Ruvettus type in the Fuller collection is figured by Beasley (4, pp. 117, 118, pl. CLXXIII). It was collected by W. W. Gill and sold by his widow; and, as it was the only object retained by Gill, it must have had some special significance to him. It is nearly two feet long, with page 242lateral projections, and carving. The large size and the weak snood indicate that the hook was made for some ceremonial purpose rather than for fishing (fig. 153, b). Though Beasley attributed the hook to Rarotonga, there is a possibility that it came from one of the other islands.

A wooden shark hook in the British Museum is also figured by Beasley (4, p. 117, pl. CLXXV) and rightly regarded by him as one of the finest hooks in existence. It is of the same shape as the shark hooks, but from below the shank knob it is carved in transverse narrow panels with the pattern characteristic of Mangaia. It has no attached snood, and it is improbable that the owner of such an exquisitely carved article would risk its loss or damage in fishing. Its fine workmanship and the fact that it originally formed part of the London Missionary Society's collection attest its age (fig. 154).

Figure 154.—Carved wooden hook from Mangaia (British Mus.): a, heavy dark wood; length, 287 mm. (11.5 inches); width, 165 mm. (6.5 inches); well formed shank knob (1) and well curved point (2); typical Mangaian carving in transverse panels commences below shank knob and continues to point, panels completely covered in original; short angular panels (3) on either side of bend to balance carving technique. b, enlargement of bend to show up angular panels (3).

Figure 154.—Carved wooden hook from Mangaia (British Mus.): a, heavy dark wood; length, 287 mm. (11.5 inches); width, 165 mm. (6.5 inches); well formed shank knob (1) and well curved point (2); typical Mangaian carving in transverse panels commences below shank knob and continues to point, panels completely covered in original; short angular panels (3) on either side of bend to balance carving technique. b, enlargement of bend to show up angular panels (3).

There is no definite information regarding the use of these hooks but they were possibly symbols retained by chiefs or head fishermen to bring success in fishing.

Bonito hooks

A paragraph is given to these widely distributed hooks because Beasley (4, pp. 36, 37), after remarking that he had searched in collections for Cook Islands specimens without success, states that Mr. E. L. Gruning informed him that he had seen bonito hooks with pearl-shell points on Manuae. These hooks were probably brought in from Manihiki, Rakahanga, or Tongareva where the points are made of pearl shell. Other Polynesian islands use bone or turtle-shell points.

page 243

The inhabitants of the volcanic islands of the Cook group state that they never fished for bonito and hence had no bonito hooks. Manihikians living in Rarotonga fished for bonito with their own hooks. It is probable that the Cook Islands were too far south for bonito schools to visit them frequently enough to excite the interest of the Cook Islanders, though their ancestors must have been acquainted with this form of fishing in the Society Islands.

Angling Methods in Mauke

My informants in Mauke described five different methods of angling as follows:

1.E a'o te tautai (a line is the fishing). Two wooden hooks on a wooden spreader were baited with octopus for sea fishing. Though a line (a'o) was used in all the methods, the use of the term a'o evidently arbitrarily distinguished this general form of fishing. The fish caught were 'apuku, mai'e, urua, ru'i, angamea, rot, tarao, and 'oka.
2.E 'oe te tautai (paddling is the fishing). The term paddle ('oe) was used because the paddle was constantly plied while the line, about five fathoms long, was trolled. The hook was wooden, the bait was koperu, and the time was day. The fish caught were mango (shark), urua, angamea, ono, varu, 'a'ai (albicore), and kaku.
3.E tutau te tautai (anchored is the fishing). The term tutau means an anchor and conveys the meaning that the canoe was anchored during operations. A wooden hook was used with a line about 12 fathoms long and without a sinker. The bait was koperu and the fish caught were the same as those caught by the trolling method. Fishing took place at night on the 'Oata (3d) night and the ni ghts of the full moon.
4.E tutau 'a'ai (anchored for albicore). This differed from the third method only in that it took place in the daytime. In addition to albicore ('a'ai), the same fish were caught as those procured by the trolling method.
5.E 'i koperu te tautai (line fishing for koperu is the fishing). The koperu were fished from the outer edge of the reef with a two-branched ironwood rod (matira) with a Look attached to each branch. The time was about half past four in the afternoon.

The koperu, a species of mackerel, was an important fish in Atiu and Mauke. The fish came in schools right up to the edge of the reef, and the parts of the reef frequented by the mackerel were owned by various chiefs. These mackerel grounds, termed manava, extended around the island, and each had its place name. The ariki Tamuera of Mauke gave me a list of 54 such grounds which commenced with Tiringa-maoukura and ended with Avaavaroa. Each had its owner, and ownership passed by descent. Some of the Mauke grounds were owned by chiefs in Atiu. I was told in Atiu that the chief Maungapu owned the Te Kopu-o-aki, Vai-to'ora, Taeta, and Takiore mackerel grounds in Mauke. In Tamuera's list, these four grounds are close together.

The fish were fed regularly in the afternoon at about half past four with cooked taro, sweet potato, banana, and grated coconut carried down in a wooden bowl. The caretaker threw the food into the mackerel ground from the reef, and the fish came in shoals. Tamuera said that good food had to be cooked for the fish because they were famous (ika rongonui) and protected (ika tapu). When the owner wanted some mackerel he sent an attendant to page 244his preserve who, after drawing them in with some food, caught as many as were wanted with the forked rod. The mackerel preserves of Atiu and Mauke form an interesting link in thought with the enclosed mullet fish ponds of Hawaii.

A myth on the local origin of the mackerel states that a woman named Pitako brought them with her when she came to Mauke in search of her husband. Her canoe came through the reef channel named Tukume but she died on reaching the shore. A mackerel ground near the Tukume channel was named Pitako after her. It occurs in Tamuera's list.

Fry

The fry (ika tauira) of sea fish commence to come into the lagoon in shoals in February. In Mangaia, they are scooped up with a kukuti net. Names are applied to the fry and to one or more stages of growth before they reach adult size. The following names are from Mangaia:

The pi'o fry are very white (teatea), but in three days, they darken in color and are termed para'a. The adult stage is the api (maito in other islands).

The ka'uru fry becomes the adult a'uru. Ia ngangao te ka'uru, ka tuatua'ia e a'uru (When the ka'uru becomes large, it is said to be an a'uru).

The koninonino fry becomes the adult avini also called tiovi (in Rarotonga, manini). The miro fry becomes the i'e (garfish).

The takuo fry, considered the best eating fry grows into the takuo-paparu. A whole shoal will follow their leader into a net.

Fresh-Water Fish

Fresh-water gobies (kokopu) are found in the streams and in fresh-water lakes such as Lake Tiriara in Mangaia. When the sea was too rough to permit of sea or lagoon fishing, the process of rotu or rutu kokopu for gobies was carried out in Lake Tiriara. In olden times, the shores of the lake were cleared of hibiscus trees ('au) and weeded. A party of about 40 men took part in the operations, standing in water up to the breast, this depth being termed one tukunga. An old nariki net was set with a man holding each end pole. The other men formed an oval formation (rau tangata) and worked toward the net splashing the water with both hands to drive the fish before them into the net. When the net was lifted, as many as three coconut-leaf baskets were filled. The lake was worked over, and even the deeper parts near the exit under the makatea wall Were explored. When operations ceased, the catch was distributed (tu'a) among those who had taken part.

Angling for goby (titomo kokopu) was done with gorges and with the primitive hooks made with pandanus spines. The bait was fresh-water shrimp (koura vat).

page 245

The fresh-water goby was fat in October and November, and the sign that they were ready to be caught was the new growth of turmeric (renga) which comes up in these months.

Fresh-water shrimps (koura vai) were caught by hand or, in Rarotonga, by a slip noose formed from a coconut-leaflet midrib. Children stalked the shrimps in the streams and placed the open noose behind the shrimps. As they backed into the noose, the midrib was jerked upward to close the noose. In modern times, small many-pointed spears usually made from the ribs of an old umbrella are used for spearing shrimps, particularly at night with a torch.

Fresh-water eels seem to be considered unimportant, except in Mitiaro where they are obtained in large quantities from an inland lagoon.

Payment of Experts

People who could afford to, employed the services of an expert to make nets and lines. The employer got together some fish such as koperu, cooked and mashed taro in the form of tiroimi, and a bunch of ripe bananas. These he took to an expert and, presenting the food, asked him to make his line or net. If the expert accepted, he said, "Ae, tikina te kiri'au" (Yes, get the hibiscus bark). When the task was ended, the expert took the net or line to his employer, who prepared a feast termed patu a'o (a'o, line) in which roasted pigs and bananas artificially ripened were presented to the expert in recognition of his labor. The new line or net was placed under a bowl containing coconut cream in order that it should secure food. If the expert were not properly rewarded, he went to his marae and complained (aue) to his gods, hence the line caught no fish.

Magic for New Nets and Lines

According to the ariki Tamuera of Mauke, new nets and fishing lines were given good luck by the following procedure. The owner's wife prepared taro and arrowroot in puddings known as poke taro and poke pia which were placed in a large wooden bowl. The new net or the line with a fisherman's basket of hooks was placed beneath the bowl. The fisherman gathered a number of children who had not eaten recently and were very hungry. He brought the group to his house and, indicating the bowl of food, told them to fall to. The manner in which the hungry children rushed the food and clustered around the bowl was held to indicate the manner in which fish would come to the net or the line that was beneath the bowl. As children who were not hungry and came quietly to the bowl would affect the future attitude of fish, the owner ' was careful to question them before extending an invitation.