SH (^5 €l^. ^ " /'a:>^ ^ p /- ^ t ■ ^?^?oses. These changes Avere abrupt and coincided Avith the sudden iiicrease in the run. The natural enemies Avhich left traces of their presence on the entering salmon were the AA^hite Avhales, or belugas, and the lamprey eels. Belugas Avere very much in evidence in the loAver river chan- nels during the latter part of the season, the size of their schools increasing as the salmon were running more abundantly. Un- doubtedly they were feeding on the salmon, and it is safe to 'assume that they captured and deA^oured a A^ery large proportion of those on Avhich they succeeded in closing their jaAvs. But the number of salmon Avhich appeared on the cannery floor bearing unmistakable tooth marks of the beluga was surprising. The sides Avere scored lengthwise by widely spaced lines, Avhich usually described a gentle curve, but were occasional!}'- angulated. Evidently these salmon had escaped from the A^ery jaAA^s of their pursuers. Other marks Avhich attracted uniA^^ersal attention and Avere usually mistaken for hatchery brands Avere the scars made by the lamprey eel. This slender eel-shaped animal has an oval sucker-shaped mouth provided Avith roAvs of rasping teeth. By means of the sucker mouth it attaches itself to the salmon and may rasp off the skin and even deeper-lying tissues for food. A scar is left which often reproduces with great fidelity the details of the mouth, with its outer fringe of filaments and its inner groups of teeth, AAdiich in the scar often giA''e the impression of printed characters. Such lamprey scars haA^e been occasionally obserA^ed in other rivers, but never before in such abundance as on the Yukon. This fact probably stands related to the large lamprey run which is indigenous to this stream. They enter the mouth of the river in the fall after the surface has frozen and run up under the ice, to the under surface of AA^hich they often attach themselves Avhen resting. Although the run lasts but few hours at any locality, it is of enor- mous dimensions and furnishes tons of food to those AAdio dip them up through holes cut in the ice. Like the salmon, the lampreys enter the riA^er for purposes of propagation, and all die after the eggs are laid. The 7/oung soon V. S. BUREAU OF FISHERIES. 133 after hatching burrow in the mud of the river banks and live like earthworms for an indefinite period of two or more years. After this they pass out to sea when some 6 inches long, and spend the remainder of their life in the ocean. So far as known to the writers, the lamprey scars have been found on salmon exclusively, and on the Yukon never on chums or dog salmon. Some king salmon would have two or even three scars and one was found on a sockeye, but among the thousands of Yukon chums that we in- spected there was not one that had been attacked by a lamprey. It is an interesting question whether the Yukon lampreys follow the king salmon of their own river on their feeding grounds and prey on them. It is not known that their attacks are ever dan- gerous. The salmon observed seemed in no case to have been seriously injured. That the king salmon ascend the Yukon at a high rate of speed has been accepted generally. An attempt was made to secure re- liable records of their first appearance at a large number of localities along the river. Wireless messages were sent to a number of points during the early days of the run before the dates should be for- gotten. And, in addition, a number of important records were secured during visits to the fishing camps, some of these giving the catch in detail da,y by day throughout the season. While it is recognized that the capture of the first salmon of the season at different points along the river may vary within a day or two in relation to the beginning of the run, an examination of the data indicates that this source of error is not serious and that reliable conclusions concerning the rate of travel can be drawn from the table presented. In this table, when two or more records have been obtained from the same locality, the earliest has been selected as giving the first appearance of the king salmon in that portion of the river. Date of Capture of First King Salmon at Localities on Yukon Rivee, Season OF 1920. Locality. Approxi- mate distance traveled . South mouth of river Run begins south mouth Pilot Station Marshall Russian Mission Tucker's fish camp Paimiut Holy Cross Halls Rapids, above Anvik Camp 51 miles below Kaltag Kaltag Koyuiuk Whisky Creek, above Louden Ruby Tanana Fish Creek, above Rampart Rapids . Circle Charlie Creek Eagle De Wolf's flsh camp Dawson June 13 June 15 June 20 . ..do.. .. June 21 June 23 June 22 June 23 June 24 June 27 June 28 June 29 June 27 ...do.. .. June 28 July 3 July 11 July July July Miles. .do.. 107 144 204 221 259 279 346 440 491 555 622 659 804 851 1,227 1,317 1,402 1,478 1,504 134 U. S. BUEEAU OF FISHERIES. Inspection of the above table shows the slowest rate of travel in the first hundred miles of the river above its mouth. From the en- trance of the South Mouth to Pilot Station is approximately 130 miles, but it was five days after the run began in the mouth of the river before the first king salmon appeared at Pilot Station, indi- cating a rate of about 30 miles per day. Between Pilot Station and Tanana, on the other hand, the rate of travel was slightly more than 80 miles per day. This discrepancy is probably due to the habit of playing back and forth in brackish water, on entering the river mouth, before beginning their serious ascent of the river. It is well known to the fishermen at the mouth of the river that salmon enter the gill nets as numerously from the upstream as from the down- stream side. They pass back and forth on the tides, lingering within the fishing district, thus giving the nets many more opportunities to capture them than would be the case if they pursued a direct course on entering the stream. Three records below Tanana, those of the camp 51 miles below Kaltag, Kaltag itself, and Koyukuk, do not align themselves with the remainder of the series. At the average rate of travel king salmon should have reached Kaltag by the 24th instead of the 28th and Koyukuk by the 25th instead of the 29th. In both of these localities the capture of king salmon was considered of little relative importance, and the records are doubtless defective. Above Tanana the current of the river increases materially, rapids are encountered, and the intricate channels of the Yukon Flats are to be threaded. It is not surprising to find that the rate of travel in the upper portion of the river becomes reduced. Not only are the dilRculties of ascent increased but the potential store of energy in the fish approaches exhaustion. When they enter the mouth of the river they are the richest in oil of any salmon known, but by the time they reach Dawson their flesh is comparatively dry and flavor- less, the oil having been expended to supply the energy needed in ascending 1,500 miles against the current and in carrying forward at the same time the sexual changes which precede the act of spawn- ing. The average rate of travel from Tanana to Dawson was slightly less than 45 miles per day, while from Pilot Station to Dawson, involving practically the entire length of the river below Dawson, the average rate was 57 miles per day. No record of any other river approaches this in completeness nor in the high rate of travel indicated. The unexampled speed Avith which salmon ascend the Yukon is doubtless associated with the great distances to be traversed before reaching their upper spawning areas, taken in connection with the shortness of the northern summer. Inasmuch as the investigators were compelled to restrict their attention to the main river, they are unable to designate the prin- cipal spawning areas of the king salmon. Limited numbers of kings are reported to turn aside into all the principal tributaries of the lower and middle sections of the river, but it is believed that a rela- tively large proportion of the run passes beyond the mouth of the Porcupine into the upper portion of the basin. CHUM OR DOG SALMON. Although the king salmon is an important source of food to the natives and the white population, it is far surpassed in value by the U. S. BUREAU OF FISHEEIES. 135 chum or dog salmon, which must be considered the principal food product of tiie Yukon River. It makes its appearance off the mouth of the river only a few days later than the advent of the king salmon. In 1920 the first chum was obtained June 17, and from the ITth to the 19th, 138 became entangled in the coarse mesh of the king salmon nets set outside the mouth of the river; but none apparently had as yet entered the stream. On June 20, 2 were reported inside the river ; June 21, 82 ; and June 22, 26; but during these three days 856 were taken in the outside nets, indicating clearly that the chums were beginning to school in some abundance outside the river, but that few were entering up to June 22, when the run up the river may be said to have begun. It became greatlj^ accelerated on June 29 and maintained itself with minor fluctuations, until the cannery ceased oi^erations the middle of July. No strictly quantitative results concerning the run of chums could be obtained from the cannery records during this season, as nets with appropriate mesh for capture of chums were not employed until the last week in June. For this reason, the apparent increase in size of run during the latter days of June is certainly overemphasized by the cannery statistics. The first chums to arrive were further advanced toward sexual maturity than were the king salmon. In the latter the testes and ovaries were small and the snout so little produced in the males that there was no certain external evidence of the sex of the individuals examined. But in the case of the chums, even those earliest to run had the milt white in color and ob^dously enlarged, and the large eggs were already loosening in the ovary. The jaws of the males were not hooked, but were showing a slight sharpening and elongation, so that sex determination could usually be made from the appearance of the head. All the early chums were bright silvery in color, with abun- dant oil, and pinkish flesh which turned a deeper red on drying. All of them were in such condition that they would have been classed as " silvers " rather than as " dog salmon " by fishermen of the upper river. But changes in the appearances of the chums were soon apparent. At first, as in the case of the Idng salmon, rare individuals, usually males, showed themselves in an advanced stage of development, with brightly colored bars on the sides of the body and long hooked jaws. They stood out conspicuously from their fellows, which were still in the " silver " stage. But by the last of June, when a great increase suddenly occurred in the take of chum salmon, obvious seasonal changes had appeared in this species, as in the king salmon running at the same period. It was now the rule for the males to exhibit elongated jaws provided with canine teeth, and to show the beginnings of the conspicuous color marks which characterize the spawning males of this species. During the first days of July the run of chums again fell off, accom- panied by a further increase in the matured appearance of the fish. As this occurred at the same time in chums and in king sahnon, it appeared to the fishermen that the end of the run was in sight. On July 7 a new run of chums suddenly set in, the great majority of .which were of bright silvery appearance and with no further sea- sonal advancement than had been shown by the fish that ran first in June. Among them were a few stragglers of the previous run, the 75376—21- — 2 136 U, S. BUREAU OF FISHERIES. males and females equally conspicuous among their silvery compan- ions. On subsequent dates these stragglers became more rare, and when the party left for the upriver on August 2 the few chums that were then running were almost wholly of the pronounced silvery type. It is unfortunate that confusion should have arisen in the Yukon from a failure to recognize that the clog salmon and the so-called " silvers " represent different phases in the development of one and the same species. The "dog salmon" are the individuals furthest advanced toward spawning. They exhibit the elongated hooked jaws and enlarged teeth in the male, the bright nuptial coloration, and the impover- ished condition of the flesh, which is light in color, largely devoid of oil, and possessing very little substance when dried. In the "silvers" the eggs and milt are less developed, the jaws of the male are little or not at all hooked, the external coloration is sil- very, or with a light flush of red, and the meat is red in color when dried, rich in oil, and valuable both for human food and for dogs. In general, the " dog salmon " along any stretch of the river con- sist of those individuals which will turn into some adjacent tributary to spawn, while the " silvers " are on their way to the upper reaches of the river, show relatively little of the sexual changes they will ex- hibit on their spawning beds, and are still richly provided with the oil which serves as fuel and principal source of nourishment during the long journey still before them. Many fishermen recognize the difficulty of distinguishing sharply between " dogs " and " silvers " and relieve their embarrassment by recognizing a third class, the " half-breeds." But the term " silver salmon " has acquired a fairly definite and useful significance in the trade. "Dog salmon" are so poor in nourishment that they have indifferent value even for dog feed, and will not be purchased except during times of extraordinary scarcity. The natives will feed them to their dogs, but will not eat them themselves unless king salmon and " silver salmon " are unobtainable. The use of the term " silver salmon " for bright silvery chums, still rich in substance, is 60 universal and of such long standing in the interior of Alaska that it seems useless to attempt to supplant it with any other name. Confusion will inevitably result owing to the pres- ence of the totally different species, the coho, which is commonly known in outside waters as the silver salmon. It is proposed, there- fore, that the term " silvers," when referring to the Yukon basin, be restricted to the chums known commercially by that name, while the three species of salmon of importance on that stream be known as the king, the chum, and the coho. It became. evident, as we were ascending the river in August, that the "dogs" and the "silvers" were in general keeping apart from each other and were following distinct migration routes. Through- out the entire lower course of the Yukon, from Tanana at least as far as Anvik, the " dogs " predominated on the right (north limit) of the river and the " silvers " on the left limit. This is generally recognized by all the fishermen of that region, who also agree that the " dog salmon " turn into all the creeks and smaller tributaries, while the silvers " dislike the taste of fresh water," as a native fisherman stated the case. It is also recognized that a heavier run of kings and U. S. BUREAU OF FISHERIES. 137 of colios is found in company with the " silvers " along the left limit of the river. It appears, therefore, that there is a prevailing use of the left shore hj those fish which are bound for the upper reaches of the river. This may have connection with the fact that the ma- jority of the tributaries of the lower river enter on the right bank. The run of chums is not of uniform character and quality through- out the season. Fishermen recognize a succession of phases in the run, characterized by fish which on the average are recognizably different. ' In general, it is stated that the run of chums which accom- pany the king salmon are 'of relatively small size and poor quality, commonly known as dog salmon. Following these is a run of bright fish of good quality but inferior size, known as " silvers," while the last chums to run, late in August and early in September, are the finest of all, the " silvers " par excellence, a bright rich form dis- tinctly of larger size. This last run of " silvers " ordinarily is of short duration, but is frequently of great intensity, and furnishes the most highly i^rizecl fish of the season. As it was necessary to make a canvass of the entire river before the fishing season had come to a close, the writers were unable to observe the alleged succession of forms of the chum salmon at any locality. Distinct differences in the character and quality of the fish appeared at the various camps visited, but these differences were apparently dependent either on the proximity of the camp to important tributaries into which spawning chum salmon would pass, or, more especially, on the location of the camp with reference to the river itself, whether on the right bank or the left. In general when following up the right (north) bank of the river the quality of the fish became poorer as one approached the mouths of the tributaries. The impression was strong that the fish destined for these tributaries vrere farther advanced than those with which they were associated on the same side of the river bound for more dis- tant spawning grounds. A further more detailed s^udy of the char- acteristics of the chums at different times and in different localities and the distribution of the various strains to their respective spawn- ing areas would offer results of importance equally from the bio- logical and from the strictly practical point of view. During the season of 1919 the bright chums, or " silvers," were said to be almost wholly wanting. If these were the fish bound for the more distant spawning beds, then the upper sections of the river must have re- mained largely unseeded in 1919 and incapable of producing their quota of a subsequent season's supply. In 1920 the run of chums was peculiar, in tha^ it lasted lono-er than is usual in good years, but was of less intensity, and the August run of " silvers " failed to attain its usual proportions. In subsequent years the late run of " silvers " should be carefully noted. If there is evidence of a pro- gessive decline, steps should be taken to protect this most important part of the run. To accomplish this result information should be obtained concerning the spawning beds which are resorted to by this large and valuable race of chums. The chums travel up the river at a rate approximately equal to that of the king salmon. In 1920, as has been shown, they were schooling on the tide-flats off the mouth of the river nearly a week before they began to ascend the stream in a.Tij numbers on June 22. Adopting the latter date as the beginning ,of the run, it appears that they 138 U. S. BUEEAU OF FISHERIES. started about a week later than the king salmon. In the section of the river between Louden and Tanana, they were reported as 8 to 10 days behind the kings, and in the district between Circle and Dawson, the first chum was recorded 11 to 17 days later than the first king. It is evident, therefore, that the chum salmon, also, although apparently far less vigorous than the king, ascend the river at an astonishing rate of speed, maintained without cessation for well over a month, under the necessities imposed on them by the short summer and the rigorous fall climate of the Yukon. The opinion has already been recorded that the king salmon of the Yukon is the richest in oil of any known king salmon. The same statement, made with even greater emphasis, may be advanced re- garding the Yukon chums. This species is noted in other localities for its poor oil and poor color. The canned product is the reverse of attractive in color or in richness, and occupies the lowest rank in the markets of the world. In other river basins the chums do not travel far from the sea, but enter late in the season and seek spawning beds not far inland. But the best of the Yukon chums travel 1,000 to 2,000 miles up a river known for its consistently rapid current. They form a rich table fish in the lower section of the Yukon, where the king salmon, to ordinary palates, contains a superabundance of oil. No more striking evidence of the richness of the king salmon can be found than that presented in the smokehouses of natives in the lower river. Here the drying salmon constant!}^ drip a pure, clear red oil, which is collected in vessels and preserved for winter use. At the village of Ohogamute. above Marshall, we observed some 30 quart bottles filled with red salmon oil, secured this season from drying king salmon. YUKON DELTA, ITS PRINCIPAL CHANNELS, AND ROUTES MAINLY FOLLOWED BY SALMON. The Yukon River subdivides numerously within its extensive delta, and sends its waters to the sea through a large number of distinct channels. Many of these are narrow winding canals, which meander in every direction interminably through the flat lands, before they attain the outer edge of the delta. Through all of them, doubtless, some salmon pass, but three of the channels so far transcend the others in size and importance that for practical pur- poses they may be considered as forming the mouths of the Yukon and the main migration routes of the entering salmon. The channels in question, in order of importance, are Kwikluak Pass, or south mouth, with its important branch, the Kwiguk Pass; Kawanak and Kwikpak Passes, which coalesce before entering the sea and form together the middle mouth ; and Apoon Pass, or north mouth. Much the smallest of the three is the Apoon Channel, which traverses the northern sector of the delta, and forms the most direct route for river steamers bound to and from St. Michael. Although exclusively used as a steamer channel, it is poorly adapted for this purpose, being so shallow both within and without the mouth that light-draft boats have quite the habit of going aground there and waiting until a favoring tide shall float them. Few salmon apparently find their way through Apoon Pass. A. few families of natives from Kotlik villao;e had established a fish U. S. B.UEEAU OF FISHEKIES. 139 camp on the shore just outside the mouth of the river, but the drying frames were scantily supplied with salmon when they were seen in the last week of July. Between the mouth of the pass and old Fort Hamilton no fishing stations were occupied, and the nets operated by the white trader at old Fort Hamilton were having no success. The general belief that Apoon Pass is of little importance as a fishway seemed wholly justified by observations. It has probably no greater importance than have some of the subsidiary channels through the delta. The lower Kwikpak and the Kawanak Passes, which together con- stitute the middle mouth, are the least known of the^ three main divisions of the river. ISTo steamers traverse them and very few natives have their summer fishing camps along their banks. The Kawanak is a stream of large size and fair depth of water and the lower Kwikpak, although choked with sandbars, carries a consider- able current. This middle mouth was visited July 5 to T, at a time when the Carlisle Packing Co., on request, was testing the run of salmon by setting nets in the lower Kawanak Channel. Two nets were set along the left bank and two others were set offshore along the edge of a bank. This test was made during a slack period in the run in the south mouth, when the king salmon especially were running in greatly reduced numbers and the chums were not coming in full force. During 14 hours' fishing in the Kawanak Channel the four nets took 3 king salmon and 67 chums. In a second test of equal length the following day the showing was even less favorable. One native fishing camp, which obviously has been occupied for many years, is located on the upper point of the long island which separates the Kawanak and Kwipak channels, immediately below their first confluence near the mouth. Four families were encamped at this place, and reported a favorable catch of king salmon during the preceding two weeks. The run had now slackened, they said, and the chums were just beginning to appear. They had found the season thus far very much better than the preceding year, when they had fished in the same locality. During the season of 1919 they had been unable to secure many more salmon than they had needed for their summer's use. At the time the camp was visited they had caught enough king salmon to fill one rack and two smokehouses, and had made use of two short gill nets of their own make set in an eddy along the bank of the island. The nets were not more than 25 feet long. Later, when a fresh run had entered the south mouth, word was re- ceived that the native fishermen in the middle mouth were again making good catches. It seemed, therefore, that the runs in the two mouths were well synchronized, the fluctuations during different seasons and between different days of the same season, following- each other closely. This was well shown in a test made in the middle mouth, on request, during the earlier part of the season, when king salmon were running abundantly in the south mouth. This test was made on June 25 and 26, and resulted in a satisfactory catch of king salmon, with very few chums. In view of the tests here indicated and observations at the fishing camp, it can not be said that the middle mouth lacks importance as a route for salmon. Yet it is considered to be very far indeed behind the south mouth in this respect. It is doubtful whether it equals in 75376—21 3 140 U. S. BUREAU OF FISHERIES. importance the subsidiary channel known as the Kwiguk, which branches off from the Kwikluak Pass, a few miles above its mouth. It may safely be inferred that the native fish camps are located at the most favorable fishing sites. These are occupied year after year and generation after generation b}'- the same families and their de- scendants. Where native fishing villages are most thickly grouped will be found the largest and most unfailing supply of salmon. With this as a basis, the writers are compelled to conclude that the south mouth serves as the migration route for the greater part of the Yukon run. While no data are available for an estimate, it is not considered beyond the bounds of probability that nine-tenths of the entire run enter by the Kwikluak Pass and its subsidiary channel, the Kwiguk. About 100 families of natives were fishing in this district in 1920, while not to exceed 10 families were seen in the middle mouth below Dogfish Village. In no case did the latter families have equal success with those camped on the Kwikluak Pass. It is on the KAviguk Channel, just below its emergence from the Kwikluak, that the floating cannery of the Carlisle Packing Co. has been located during the two seasons of its operation in the delta. Protected from the heavy southerly- winds which blow up the main channel during the summer months, this site is within easy distance of the main fishing grounds in the lower part of the Kwikluak Chan- nel and among the offshore shoals and islands. Inasmuch as the Yukon salmon appear largely to travel along the banks, in the eddies, and along the margins of submerged banks, it is believed possible during a favorable fivshing season to secure from this location as a base a very considerable proportion of the salmon of the Kwikluak Channel. Fortunately, no commercial fishing for export is per- mitted in the Kwiguk Channel. Several native families were lo- cated on this channel in 1920, and one white trader maintained a wheel. Good catches of king salmon and chums were secured at all of these camps, but it was believed that the run of kings was proportionally^ not as heavy as in the main channel. COMMERCIAL FISHERY OPERATIONS IN YUKON DELTA IN 1920. The fishery operations of the Carlisle Packing Co. in 1920 were conducted principally in the south or Kwikluak mouth of the river and beyond that mouth among the seaward channels which diverge from it. The mouth of the Kwikluak Channel was designated as it was during the previous season by a stake set in the right or north- ern bank on the projecting point of land at Ingrakaklak (see U. S. C. & G. S. chart 9373) and by a stake set on the left or southern bank of the channel at the entrance to the well-marked lagoon some half mile below Nilak. A few fishing camps were established by the company as far up the Kwikluak Channel as Dogfish Village, where the main river makes its first grand division into the Kwikluak Channel and a channel which later divides to reach the middle and the Apoon mouths. Above Dogfish Village no fishing camps were established by the company, but they purchased limited numbers of salmon from a few independent fishermen, who operated at points below the mouth of Clear River near Andreafski. U. S. BUREAU OF FISHERIES. 141 Commercial fishing for export was thus carried on exclusively in the main Yukon, between the mouth of Clear River and Dogfish Village and below Dogfish Village, in the main Kwikluak Channel and its seaward extensions. All subsidiary channels branching off from the Kwikluak were closed to commercial fishing, whether these served as communicating links between larger channels, as in the case of Aproka Pass, or, as in the case of the Kwiguk Channel, secured independent egress to the sea. But in no case were the salmon which entered through these pro- tected channels given unimpeded access to the upper river. All of them must pass through the 40 or 50 mile stretch of the main river between Dogfish Village and Andreafski, where they were exposed to capture for commercial purposes; and those that enter through the important Kwiguk Channel must in addition run the gauntlet of a further 40 miles of river between the Kwiguk entrance and Dogfish Village. During the past season very little commercial fishing was in fact carried on in the stretch of river above Dogfish Village, but this was of choice and not from necessity, for no restrictions are there imposed by existing regulations. During the season of 1919, 65 per cent of the king salmon and 61 per cent of the smaller fish — chums, cohos, and sockeyes — were taken beyond the mouth of the river, while in 1920 the proportion was even greater, 69 per cent of the king salmon and 68 per cent of the chums being taken outside. The fishing grounds in 1920 extended much farther away from the river than in 1919. During a part of the season 12 fishing boats were located between 10 and 20 miles outside Nilak, along the shallow banks bordering the Acharon Channel on the mainland side (U. S. C. & G. S. chart 9373). The outermost stations were for a time the most successful, meeting the Yukon salmon well down the coast toward the mouth of Black River. How much farther in the direc- tion of the Kuskokwim the advancing schools may be encountered in numbers adequate to warrant commercial fishing is as yet undeter- mined. An expedition which they sent down the coast to Cape Romanof testifies to the interest of the Carlisle Packing Co. in this question. The fishing methods employed by the company in 1920 did not differ from those in use in 1919. Almost their sole dependence was on gill nets, set in convenient lengths in the eddies and on shallow banks along the main channels. For the most part these were set nets or anchored gill nets, with one end made fast to the shore and the outer end anchored. But during the latter end of the season, when the river was no longer at flood, it became possible to fish on shallow banks, which were not available during the height of the king salmon run. Stake nets were then used in larger numbers and would unquestionably have been availed of more extensively throughout the season had the stage of water permitted. The gill nets used were of two kinds, the king salmon nets of 8|- i.nch mesh and the nets for chums and other small salmon of 5f-inch mesh. As the company from the beginning of the season contem- plated fishing largely, if not exclusively, for the king salmon, only nets of the larger mesh were issued during the height of the king salmon run, which lasted up to the last days of June. After this 142 U. S. BUEEAU OF FISHERIES. date the smaller mesh nets were also issued, but the number of these on hand was not adequate to supply all the fishermen. Forty-one boats in all were employed in fishing, 34 of these throughout the season, the remainder for varying periods. Four of them were engaged so short a time and obtained so few fish that they may well be omitted from consideration. To each boat was issued 200 fathoms of king salmon net, and subsequently 200 fathoms of the small mesh salmon net, to the extent that this was obtainable. In addition to the gill nets, renewed attempts were made in 1920 to use fish wheels. Two of these were constructed by the company, but were no sooner placed for fishing than they Avere put out of commission by the heavy drift which was running during the early days of the fishing season. In addition to the above, nine wheels were privately owned and operated within the area open to commercial fishing, and the catch of seven of these was sold in whole or in part to the cannery. A very limited number of salmon were obtained from independent fishermen using nets. The number thus purchased from independent operators using wheels or nets amounted to some 7,400 kings and 27,000 chums, out of a total number handled by the cannery of 58,467 kings and 155,655 chums. The fishing gear operated by tHe cannery and by private parties occupied only a narrow fringe along the margins of the channels. No attempt was made to fence or block the main channels in any manner, nor could such an attempt at any time be successful. The employees at the cannery numbered 254. Of these, 122 were brought in from Seattle, including 40 Orientals and 44 fishermen. Of the 132 Alaskans, 40 were employed as fishermen, while 48 were natives, of which the first crew of 25 ceased work in the middle of the season. It was pleasing to learn from the superintendent that the Alaskan fishermen gave a good account of themselves and would hereafter be preferred by the company. The pack put up by the cannery in 1920 was far less extensive than in the previous 3^ear, as is shown by the following table : Pack of Salmon by Yukon Cannery in 1919 and 1920. Product. 1919 1920 Canned (cases): 28, 582 24, 543 3,181 28 15, 934 Chums. . . . . ... 12, 819 Cohos Reds 5 Mild-cure (tierces) 145 i The comparative lack of success in 1920 was due in part to the un- favorable fishing conditions and in part to voluntary shortening of the fishing season. In 1919 fishing was continued until the close of August, but in 1920 only until the middle of July, as the low market price for chums did not warrant the company in continuing to oper- ate after the run of king salmon was over. The season of 1920 was extremely late on the Yukon, and the king salmon entered with a rush, while yet the river was very high and was carrying down enor- U. S. BUKEAU OF FISHERIES. 143 mous quantities of drift, which clogged the nets and rendered them inefficient. Also, during July the weather was unusually stormy, in- terfering with the fishing and endangering both fishing gear and the lives of the fishermen exposed on the fiats beyond the river's mouth. From the cannery standpoint, the season was a failure, redeemed from actual loss — if such indeed was the case — by the extremely high price quoted this year for king salmon. Yet the cannery pack of 28,758 cases fell short only 1,242 cases of the maximum number of salmon which had been designated by the Secretary of Commerce as safely to be spared for commercial purposes from the Yukon River run. No fish intended for export from Alaska were salted on the Yukon River in 1920. SUPPLY OF DRIED SALMON PREPARED ON YUKON RIVER IN 1920. On the voyage down the Yukon from Lower Lebarge to the delta, following close behind the running ice. May 24 to June 13, few indi- cations could be observed of preparation for the fishing season. Here and there a white fisherman was engaged in constructing his fish wheel in time for the short king salmon run, but the native fish camps were unoccupied. Later it was evident that in comparatively few instances did the natives have wheels in the water in time to obtain any considerable number of king salmon. On the lower river, below Holy Cross or Paimiut, the natives be- long to the Innuit stock, and fish much less extensively with wheels than do the Indians of the upper river... They employ for the most part short lengths of liomemade gill nets, which they set in eddies behind projecting points of the shore. As favorable localities are found almost exclusivel}^ along the high right (north) bank of the lower river, the fishing villages are confined to that side. Above Holy Cross the use of nets becomes less and less an im- portant factor, and wheels are relied on almost exclusively for the capture of salmon. Rarely was the primitive fish trap or basket or the dip net seen in use. The small fish wheels, which seem to have been introduced on the Tanana River in 1904, ha vie been generally adopted on the upper river by whites and natives alike. They cost about $50 each, in addition to the labor of building them, and are wonderfully effective Avhen skillfully placed. In the section of the river between Holy Cross and Rampart little dependence is placed by the natives on the king salmon. There is an early short run, and the natives are traditionally dilatory in mak- ing preparations. By the time their wheels are in the water the king salmon run is largely over. It is also true that the king salmon are more difficult to preserve, being larger in size and richer in oil. Those that are put up' by the natives are kept largely for their own consumption and for this purpose are most highly prized. Taking the river as a whole, a distinct hardship is imposecl on whites and natives alike when the king salmon run is beloAv normal. Unquestionably, however, the chum furnishes by far the larger share of the dried salmon. Along some stretches of the river almost complete dependence is placed on this species, locally known as the dog salmon and the " silvers." The higher grade of chums, known as " silvers," form the staple dog food throughout the Yukon country. 144 U. S. BUEEAU OF FISHERIES. All the traders handle them and may deal in from 5 to 50 tons in a year. But they refuse to purchase dog salmon except as a last re- sort. The majority of the natives at the close of the fishing season sell a portion of their salmon supplj^^ to the trader with whom they deal, frequently leaving themselves without adequate provision for their families and their dogs. Later in the year they are often com- l^elled to repurchase dried salmon at an advanced price, paying for it with the proceeds of their winter trapping. They are, of course, more or less improvident, as in the case of other primitive peoples. Their sale of salmon in the fall is frequently to liquidate their debts to traders who had extended them credit eatlier in the season. In the section of the main river below Kampart, where salmon are still rich in oil and the rainfall during the summer months is usually heavj^, resort is had to smoking the salmon in order to preserve them. There is no commoner sight along the Yukon than the cluster of white tents in some picturesque nook among the hills of the right bank, and with them one or more high, barnlike smokehouses, which emit a faint blue vapor. There will be a fish wheel turning in the current along the rocky shore and a number of open-air racks, more or less protected from the weather, on which the salmon are hung for a time until partially dried and ready to be smoked. The pic- ture is, of course, not complete without the native men, women, and children of the summer camp, nor without the invariable row of dogs closely tethered to stakes driven near the water's edge. Here the dogs fatten on the salmon heads and back bones and other refuse. They scratch out shallow holes to lie in alongside their stakes or burrow deep into the adjacent bank, if one be at hand, to escape the implacable swarm of mosquitoes. Along the Tanana and the upper Yukon is a region of less rain- fall, in which also the salmon have relatively dry meat, which is easily preserved. Here smoking is frequently dispensed with and dependence had entirely on air drying. But, by whatever method prepared, the fish of the upper river, of the Innoko, the Kojmkuk, and the Tanana, are of inferior' grade, and bring a lower price than do fish imported into these districts from the main river. The best product of all is secured from the Rampart Eapids. Here the " silvers " are said to average larger and fatter than in any other sec- tion. It is not improbable that inferior strains of dogs and " silvers " have turned into the lower tributaries, leaving at the rapids almost exclusively high-grade fish bound far up the river. In the coastal district when salmon are running abundantly trenches are often dug in the soil by natives and hundreds of salmon are thrown in without preparation of any kind. They are then covered with earth and nature is permitted to have her unrestricted way with them. When the contents of these trenches are scooped out at some convenient season, perhaps in midwinter, they are said to make acceptable dog feed and to be not wholh^ shunned by the natives themselves. The king salmon intended for their own food is often carefully prepared and stared away by natives of the lower river. When suffi- ciently dried and smoked, the sides are cut into pieces of convenient size and packed solidlj^ in large baskets made for the purpose of woven grass, or Avillow roots, or frequently of salmon skins which U. S. BUREAU OF FISHERIES. 145 are neatly fitted together and sewed with sinew. The dog and " sil- ver " salmon are tied in bundles weighing about 60 pounds and stored away in the caches in this shape. King salmon are also put in bun- dles on the upper river. Mention has been made of the salmon oil obtained as drippings from king salmon. In addition to this product, the eggs are very generally saved, being closely packed in any convenient receptacle, without special attempt at preservation. During the early part of the king salmon run, from June 25 to July 1, a launch trip was made from Kwiguk to Holy Cross and return for the purpose of inspecting the condition of the salmon racks and obtaining a clue to the extent of the king salmon run along the river. The run had been on in this district about one week, and all the fish racks along the river contained considerable numbers of king salmon. The opinion was general that the season was opening favorably. It is in the delta and along this stretch of river that fishing is conducted largely with short lengths of gill nets set in eddies, and in this district the run of king salmon is a very important factor in providing the winter's supply of food. In addition to the native camps, there are numerous fishing sta- tions occupied by white men. Many of these men are married to native wom^en, and some of them are found in native camps, divid- ing the proceeds with the native families. In such cases we observed commonly that the efficiency of the camp was increased. The white men operating on the largest scale were usually holders of winter mail contracts, which necessitated the use of large numbers of dogs. One such mail carrier keeps some 60 dogs and requires annually for their feed from 12 to 14 tons of dried salmon. As these are exclu- sively dog or " silver " salmon, the number used is between 15,000 and 20,000. Other white men take dogs to board during the summer and sell all salmon they are able to put up beyond their own needs. Many of the more intelligent natives now count on doing the same. There is a wide and legitimate demand for dried salmon, for use during the winter season when all travel in the interior of Alaska must be by clog team. As it is impossible to carry enough salmon for a long journey, all "dog mushers" depend on the country they pass through. The road houses maintain a supply for this purpose, the dried salmon taking the place of hay and grain in regions where horses are em- ployed, and being equally indispensable. During the month of August and the first 10 days of September the Yukon was traversed from the delta to Dawson and the Tanana as far upstream as Nenana, traveling in a launch and calling at the fish camps on the way. The number of families was ascertained as accurately as possible, the number of wheels engaged in fishing, and the total amount of dried salmon prepared for the season. In practically every native camp visited, one or more persons had a sufficient understanding of English to enable them to furnish the required information. In the majority of instances the racks, smoke- houses, and caches were inspected, and finally some facility was ac- quired in verifying by observation the estimates furnished. By this method the writers are enabled to present an estimate of the dried salmon prepared on the Yukon in 1920, which is based on 146 U. S. BUREAU or FISHERIES. more extended data than any heretofore furnished. That it is an understatement of the amount of salmon actually captured and used on the river is obvious from the following considerations : 1. The lower river was canvassed from the 1st to the 15th of August, and there was a later run of " silvers " of limited extent, of which the figures give no account. A message from Holy Cross, dated September 15, indicated a medium run of " silvers " for some three weeks after that point was passed. The still later run of cohos is also not included in tlie estimates. 2. No clue could be obtained as to the number of salmon eaten fresh during the season, but this must be a considerable item. 3. None of the tributaries of the Ytikon were visited, with the exception of the Tanana below Nenana, yet some of these, like the Innoko, the Koyukuk, the Porcupine, and the Stewart, are im- portant streams. The natives in these regions draw on the rivers for their supply of dried salmon, and the white prospectors and miners out on the creeks may obtain their -do