ATLANTIC SALMON; SEA SALMON; SILVER SALMON; BLACK SALMON; PARR; SMOLT; GRILSE; KELT
[Jordan and Evermann, 1896-1900, p. 486.]
The Atlantic salmon is a graceful fish, about one-fourth as deep as long, deepest below the dorsal fin, whence it tapers toward both head and tail; and oval in cross section. Its head is small (about one-fifth, or less of the fish's length, not counting the caudal fin), its nose is blunt, eye rather small, and its mouth gapes back to below the eye. The dorsal fin (about 11 rays) stands about midway between tip of snout and base of tail fin; the ventrals are under the rear end of the dorsal. The anal is similar in form to the dorsal but has only about 9 rays (7 to 10 have [page 122] been recorded), whereas the humpback has 14 anal rays or more. The tail is only very slightly emarginate in adults, and is almost square in large fish, but is more forked in fish that have been at sea for not more than one year ("smolts" and "grilse").
The salmon is silvery all over while in the sea, with brownish back and with numerous small black crosses and spots on head, body (chiefly above the lateral line), and fins. The young fish (or "parr") are conspicuously marked with 10 or 11 dark crossbars while in fresh water, alternating with bright red spots, much like young trout. Fish that have been at sea for only one year (grilse) are marked with a larger number of black spots than the older fish.
The largest salmon we find mentioned was an English fish of 83 pounds. None even approaching this size is recorded from our side of the Atlantic, where a 50-pounder is unusual, though fish of 40 pounds are not uncommon in some of the larger rivers emptying into the Gulf of St. Lawrence. Very few fish reach 40 pounds in the Penobscot or St. John Rivers, and 30 pounders are unusual there, the usual run being 10 to 12 pounds. Taking one river with another, large and small, 10 pounds may be set as a fair average of the mature Gulf of Maine fish. A 2 foot fish will weigh about 6 pounds, one of 3 feet, 16 to 20 pounds, with allowance for individual and seasonal variation.
The teeth and the scales afford the most certain distinction between small salmon and the New England sea trout (p. 120). In the roof of the mouth is armed both with a cluster in front and with a row of stout conical teeth running back along the mid-line, easily felt with the finger, whereas the sea trout has the forward group only. The scales of the salmon are so large that they are seen easily, whereas those of the trout are so minute that they are hardly visible. Old salmon sometimes lose the teeth on the roof of the mouth, but large size and large scales identify them at a glance.
It should also be easy to tell an Atlantic salmon from a humpback (should any of the latter still exist in our Gulf) for the black spots on the upper part of the body of the humpback and on its tail fin are more close set and much larger and conspicuous than the dark markings on a salmon. A more precise difference is that an Atlantic salmon never has more than 10 rays in its anal fin, whereas the humpback always has at least as many as 12, while most of them have 13 to 17.
The danger will be greater of confusing smallish Atlantic salmon with silver salmon, if the attempts now in progress to establish the latter in our Gulf should succeed, for the two fish look much alike. A reliable criterion is, again, the number of rays in the anal fin, for the silver salmon always has as many as 13 of these, an Atlantic salmon never more than 10.
It is no wonder that the life of the salmon has been the subject of much scientific study and that a whole literature has grown up about it. As everybody knows, the salmon lives the greater part of its life in the sea and makes most of its growth there but spawns in fresh water.
The salmon are silvery and very fat when they enter fresh-water on the spawning journey, but [page 123] they lose condition gradually as they work upstream, for they feed very little in fresh water, if at all; they make no attempt, for example, to capture the parr they meet. Most anglers believe that they may occasionally snap up a small fish or other tempting morsel. Many are caught on artificial flies, while every salmon angler knows that they will sometimes take a hook baited with angleworms or with prawns. It has been suggested that salmon recover the feeding habits of their youth to some extent after they have spent some time in the river, for they often rise to floating insects. But the stomachs of salmon caught in fresh water never contain anything but a little yellowish green fluid. And the fact that they keep better with bellies intact than if opened and gutted suggests that the secretion of effective digestive juices has ceased.
The maturing salmon of both sexes lose their silvery sheen in fresh water during the summer months, to take on a dull brownish or reddish hue, while the belly suffuses with some tint of red, large black spots develop, and the male not only becomes variously mottled and spotted with red or orange, but his jaws elongate, the lower becoming so hooked that only the tips come together. His body becomes slab-sided, his fins thicken, and his skin is covered with slime, until altogether he is but a caricature of the beautiful silvery creature that came in from the sea.
In small streams salmon may spawn only a short distance above the head of tide; but they may run upstream for more than 200 miles in large rivers that are not obstructed, as they do in the St. John system in New Brunswick. In Gulf of Maine rivers they spawn in October and early November, on sandy or gravelly bottom, the females smoothing a shallow trough or redd and covering the eggs with gravel.
As it is with the life of the salmon in the sea that we are concerned here, the reader is referred to Belding and to Kendall for recent accounts of the mating actions of the males and females. The spent fish, known as "kelts," "slinks," or "black salmon," are thin, weak, and so exhausted that many of them die. Most of those that survive in small rivers drop down at once to the sea after spawning. But many of them linger over the winter in large rivers, improving somewhat in condition and becoming more silvery, though they take little food. If they survive the winter (which many do not, for spawning leaves them thin and exhausted) they drop downstream to salt water the following spring.
The large (6 to 7 mm.) thick-shelled eggs lie loose on the bottom and develop so slowly in the low temperature of winter that hatching does not take place until late in the following April or early in May. The newly hatched larvae are 15 to 18 mm. (0.6—0.7-inch) long, and carry a very large yolk sac for about 6 weeks, hiding among the pebbles of the spawning bed and taking no food. When the yolk sac is absorbed the little fish, now known as "parr," begin to swim and feed.
Parr live in fresh water for longer or shorter periods according to locality or to other factors not well understood. In the St. John, and in the rivers of Minas Basin, most of them remain for 2 summers and 2 winters, running down to the sea the third summer. But Huntsman has found that some few stay in the Fundian rivers for 3 years. Most of the salmon of the Penobscot spend 2 years as parr, a few 3 years, according to Kendall. It is even possible that some may linger in Gulf of Maine rivers for 4, 5, or even 6 years, as is known to happen in Norway. And Dr. Huntsman informs us that some of the male parr in the rivers of the Chignecto Peninsula become sexually mature before visiting the sea.
Parr may be moving downstream any time from late spring to autumn, but most of them probably make the journey in June and July in Gulf of Maine streams, when they are 5 to 6 inches long. They put off their barred and spotted pattern as they near tidewater, to assume the silvery coat worn by the salmon during his sojourn in the sea. They are now known as "smolts."
Salmon, small or large, are voracious while in salt water, feeding altogether on live bait, chiefly on fish and on crustaceans. Among fishes available to them in this side of the Atlantic, launce, herring, alewives, smelt, capelin, small mackerel, haddock, small sculpins, and even flatfish have all been reported as entering into their diet in one place or another. Salmon caught off Norway are sometimes packed full of herring, and a hook [page 124] and line fishery is carried on for salmon in the Baltic, with herring for bait, while in British waters salmon are sometimes caught on hooks baited with launce and with pieces of mackerel. Launce and capelin had been the chief diet of thousands of salmon opened by Comeau in the northern side of the Gulf of St. Lawrence. And it is probable that the salmon of the Bay of Fundy and open Gulf of Maine feed chiefly on herring (herring up to 5 inches long have been found in salmon stomachs near Eastport) and on launce, taking alewives or any other small fish as occasion offers, including smelts and mummichogs (Fundulus), when they first enter the estuaries.
Salmon also feed greedily on euphausiid shrimps (fish entering the Penobscot have been found full of "shrimp," probably euphausiids); to some extent on pelagic amphipods (Euthemisto), while sand fleas (gammarid crustacean) are described as ranking with launce and herring as salmon food in the North and Baltic Seas. Salmon are also credited with eating crabs.
Smolts, on the other hand, fall prey to any large predaceous fish (they have been found in the stomachs of pollock), but salmon are so heavy and strong after one or two years' sojourn in salt water that only fish as large as tuna, swordfish, or the larger sharks can menace them. Their worst enemy is the harbor seal, which is a common inhabitant of the northeastern coasts of the Gulf of Maine and of the Bay of Fundy.
The young smolts grow so rapidly on the abundant diet the sea affords that they usually reach a length of at least 16 inches and a weight of anywhere from ½ to 7 pounds after one year at sea. They are now known as "grilse." And older salmon continue to put on length and weight very fast, as long as they remain in salt water. Thus, several St. John fish which were tagged and released in the river in the autumn after spawning and which were recaptured the following summer after wintering in the sea had gained 2 to 8 pounds in weight, one of them more than 6 inches in length. Others which spent two uninterrupted years in the sea (as shown by their scales) averaged about 10 pounds heavier and 6 inches longer when recaptured. But they grow much less rapidly in winter than in summer. And they hardly grow at all during the years when they spawn if they enter the river early in season, though they continue growing until later if they enter late. Hence the size of a salmon depends more on the number of times it has spawned and on the date when it enters its river than on its age.
Most of the exceptionally large fish of 40 to 50 pounds are virgin females entering fresh water for the first time, but some are fish that have already spawned once. An interesting case is that of a 45-pound 2-ounce fish, caught in the Moisie River, on the north shore of the Gulf of St. Lawrence, June 1950, by E. E. Steedman, the life history of which had been as follows: hatched spring 1942; went to sea June 1945; returned to river and spawned there in 1948; returned to sea autumn 1948; remained there until June 1950; then returned to the river, to be caught on a "Lady Amherst" fly; age 8 years.
Some salmon become "river mature" and return to spawn after only one year at sea; these, known as grilse, are distinguishable from the older fish by more forked tail, more slender body, thinner scales, and more numerous spots that are blue rather than black. Some spawn 2 or 3 years in succession, and hence never grow large; others spawn twice in alternate years; a few three times, very few oftener. It follows from this that large salmon are to be found in the sea throughout the year, though fewer of them in summer when the spawning fish are in the rivers, than in winter when the whole stock is in salt water except for the "parr," a few immature grilse (p. 129), and such of the spent fish as winter in the rivers. Some spawn only once after 3, 4, or even 5 years at sea, growing to a great size meantime. But very few salmon live to be more than 8 or 9 years old, including the time spent in fresh water as parr.
Our ignorance of the way of life of the salmon in the sea has recently been characterized as abysmal. Certainly they are swift swimmers, and the nature of the catches suggests that they [page 125] live scattered for the most part. But at least one case has come to our notice of a school seen, and some of them netted. While salmon often leap in the estuaries on their return journey and in the rivers, we have never heard of one doing so at sea. And they keep so constantly to the mid-depths that they are seldom seen at the surface, except in the estuaries. But this rule has its exceptions, for the school mentioned above was sighted at the surface, where they were mistaken for pollock. On the other hand, there is no reason to suppose that many of the Gulf of Maine salmon descend to any great depth, winter or summer. The weirs, gill nets, and other gear that yield so many in various regions, are all operated in rather shoal water (the Baltic hook-and-line-fishery is carried on at about 1½ fathoms). Dr. Huntsman informs us that salmon are taken on hand lines in mid-winter in the Bay of Fundy. They are caught occasionally on long lines in the Gulf, and otter trawlers get stray salmon on the offshore Banks (p. 126), proof that at least some may go as deep as 50 fathoms or so, while diet (p. 124) proves that they sometimes feed near bottom if not actually on it.
Coastal waters of both sides of the North Atlantic, entering rivers to spawn. On the European side its range extends northward well within the Arctic Circle; southward to the Mino River, at the boundary between Spain and Northern Portugal, perhaps with a few reaching the Duero River, midway of Portugal. It occurs in a few rivers in western Greenland. On the American side salmon ran up all suitable rivers, formerly, from northeastern Labrador to the Housatonic emptying into Long Island Sound; perhaps the Hudson also. The northern limit of the commercial fishery for it on the American side is only about latitude 54° N. (Indian Harbor, north shore of Hamilton Inlet). And while it is known to range to Hudson Strait, reports of it from stream mouths northward from Hamilton Inlet seem often to have been based on the sea run form of the Arctic charr Salvelinus alpinus, which also grows large in the sea.
When the white man first came to New England and to the Maritime Provinces, he found salmon in every large stream not barred by impassable falls, from Cape Sable to Cape Cod; i. e., in all the Nova Scotia and New Brunswick Rivers, tributary either to the open Gulf of Maine or to the Bay of Fundy, and in the following rivers in New England: St. Croix, Dennys, Orange, East Machias, Pleasant, Narraguagus, Union, Penobscot, St. George, Medomak, Sheepscot, Androscoggin, Kennebec, Royal, Presumpscot, Saco, Mausam, Piscataqua, and Merrimac. One New England river, however, after another was so obstructed by dams after the beginning of the past century, that salmon regularly entered only the St. Croix, Dennys, East Machias, Machias, Penobscot, Sheepscot, Kennebec, and Androscoggin by the 1880's. The Kennebec was still an important salmon river as late as 1895. But by 1925 the Dennys and the Penobscot alone, of the rivers of Maine, saw regular runs, with a few fish in the St. Croix where pollution by sawdust was not as bad then as it had been, perhaps with an occasional fish in other streams.
The fate of the salmon in the Merrimac typifies its history in the rivers from which it is now barred. Salmon spawned plentifully in the upper tributaries, especially in the Pemigewasset, as late as 1793 (in 1790 the run was so abundant in the lower river that 60 to 100 a day was the usual catch with a 90-yard seine near the mouth at Amesbury), but the completion of the dam at Lawrence in 1847 completely barred the upper reaches of the river. For some years thereafter salmon congregated below the Lawrence dam in spring and summer, vainly endeavoring to ascend, but there has been no run of salmon in the upper Merrimac since 1859 or 1860, when the last salmon hatched above the darn had lived its span of life, nor have any spawned there since then with the possible exception of a few that have been lifted over the dam by hand.[page 126]
Enough salmon to yield a supply of eggs for artificial hatching continued to enter the lower Merrimac up to 1893 and there seems to have been what almost might be described as a run there in 1896, when salmon were seen leaping below the Lawrence dam nearly every day from June 10th to July 25th, often 10 or 20 at a time, and a few were lifted over. But we have not learned of a single sea-run salmon seen in the Merrimac since 1901, though watch has been kept for them by the wardens of the Massachusetts Division of Fisheries and Game, and it is not likely that salmon would still run in the Penobscot were it not for the artificial propagation that is carried on there by the U. S. Fish and Wildlife Service. But the salmon situation now shows signs of improvement, for the run in the St. Croix has increased; salmon have reestablished themselves in the Narraguagus and provide sport there for many anglers since one obstructing dam has washed out and another opened. Enough salmon run regularly in the Dennys to attract anglers and a few also in the Machias and Pleasant Rivers. The Fisheries Commission of the State of Maine, and the U. S. Fish and Wildlife Service, are now concerned with the possibility of improving the runs in these streams, and of reestablishing runs of salmon in other Maine rivers.
Along the Canadian shores of our Gulf a few salmon still run in the Tusket, Salmon, and Annapolis Rivers; many in the Shubenacadie River in Nova Scotia, some in the Petitcodiac, and great numbers in the St. John River in New Brunswick, which still is a famous salmon river.
After the smolts reach salt water they are found for a time in the river mouths and about estuaries. No doubt the little salmon (too small to sell) that are caught in summer and autumn in weirs at Matinicus Island have come from the Penobscot a month or two previous. They drop out of sight in winter, as do the older and larger salmon as well. But there is no reason to suppose that many of them go far out to sea in the Gulf. Odd salmon stray, it is true, as much as 90 to 100 miles seaward off the outer coast of Nova Scotia, while otter trawlers pick up odd salmon from time to time in the South Channel, and even on Georges Bank up to 160 miles or more at sea from Cape Cod. But the great majority of the salmon that are caught in the Gulf are taken within 25 miles of the land.
The Gulf of Maine salmon also appear to continue rather closely localized as a whole, not only within the coastal belt, but within the zone of fresh-water influence from the particular rivers or river systems from which they come. So few, for example, are caught near Cape Sable that there can be no general movement around the Cape by the fish that spawn in the rivers of the outer coast of Nova Scotia. Most of the fish that go to sea via Minas Channel from the Shubenacadie, and the few from smaller streams that discharge into Minas Basin, seem to remain along the Nova Scotia shore within a distance of 30 to 40 miles to the westward. And while tagging experiments have proved that some of them scatter more widely; i. e. to Cobequid Bay, to the estuary of the St. John River, to the Annapolis Basin, and to St. Mary's Bay, few of them leave the Bay of Fundy (for some that did, see p. 127).
The much more numerous salmon from the St. John appear to hold rather closely to the tongue of low salinity that extends westerly from the mouth of the river, keeping out from the shore, for hardly any salmon are caught either on the New Brunswick shore to the eastward, except for a few near the head of the Bay (doubtless the product of the Chignecto, Bay river system) or farther west than Point Lepreau, or around Grand Manan Island which stands directly in the route of any fish moving westward out of the northern side of the Bay of Fundy. Thus it appears that a radius of, say, 40 to 50 miles would enclose the wanderings of most of the St. John River fish.
The evident failure of salmon from the St. John to follow the myriads of sardine sized herring into Passamaquoddy Bay is especially interesting. The weirs there pick up a few salmon, the presence of which can be credited to the small run in the St. Croix River. And the numbers of salmon that are caught thence westward along the coast of Maine [page 127] are not larger than can be credited to such of the Maine rivers as still have runs of salmon.
It seems certain, also, that only odd salmon from the Penobscot and from the rivers farther east ordinarily disperse westward and southward beyond Casco Bay, for while the average catch for the coast of Maine east of that point has averaged about 12,000 pounds (some 1,200 fish) for the 10 most recent years of record combined, the corresponding 10-year average for the whole western side of the Gulf from Cape Elizabeth to the elbow of Cape Cod was only 600 to 700 pounds, or some 60 to 80 fish at most, with more than 100 pounds reported in only 5 of the 10 years and none in 3 of the years. Further evidence of a more general kind that Gulf of Maine salmon do not scatter far as a rule is that they appear about the river mouths in spring so soon after the ice goes out that they cannot have come from any great distance.
A few do stray as far as Cape Cod Bay in most years; witness catches of one to 5 or 6 fish (10-55 pounds) in 14 out of 16 years by 8 traps, at North Truro, Cape Cod, during the period 1935 to 1950, in the months of May, June, July, September, and November.
A year comes from time to time when a considerable number are taken off the coast of Massachusetts. The most recently recorded instance of this sort fell in 1937, when floating traps along the North Shore of Massachusetts Bay picked up 4,400 pounds of salmon. All of these were taken close inshore. But the 1,600 or so salmon (16,050 lb.) that were reported for Massachusetts in 1928 (the big year next previous) seemingly were farther out at sea, for all of them either hooked on long lines (10,134 lb.), or were taken in otter trawls. These must have come from as far as the Penobscot, if not from the Bay of Fundy, which is equally true of the salmon that are caught around Marthas Vineyard from time to time. One, however, of about 10 pounds, reported in the North River, Marshfield, in the summer of 1938, and a few seen jumping in the Parker River (also in Massachusetts) in the summer of 1951, may have been the product of attempts to stock these streams. Occasional salmon that have been taken along the New Jersey coast and off Delaware may have been the product of attempts to stock the Hudson.
Salmon, also, of 25 to 50 pounds that are sometimes caught in Minas Channel at the head of the Bay of Fundy, must come from afar, as Dr. Huntsman points out, probably from the Gulf of St. Lawrence, there being no run of fish so heavy in any Bay of Fundy river or in any Maine river.
It is not astonishing that some salmon should stray far afield in Gulf of Maine waters, for marked salmon have been known to make much longer journeys, elsewhere. Thus fish marked in the southern side of the Gulf of St. Lawrence have been recaught on the north shore of the Gulf; in Newfoundland; and in the Strait of Belle Isle. One marked at Bonavista on the east coast of Newfoundland was retaken 98 days later in the Margaree River, Cape Breton Island, Nova Scotia, 550 miles away by the shortest possible route. One marked in Minas Channel at the head of the Bay of Fundy went out around Nova Scotia to Chedabucto Bay on the northeast, near the Gut of Canso, a journey of at least 440 miles. Five, tagged in the Annapolis River system, were recaught on the east coast of Newfoundland, a minimum distance of 900 miles, while a sixth, from the same lot, was taken at Ramah on the outer coast of Labrador, more than 1,000 miles still farther away to the northward. This last is the most spectacular case of wandering yet reported for any Gulf of Maine or Gulf of St. Lawrence salmon.
What is chiefly interesting about the large catches that are sometimes made off Massachusetts is their demonstration that so many fish may occasionally wander so far afield. And this applies not only to large salmon but to smolts in their first year at sea, for salmon so small that they must have run down to salt water but a few months previous have been taken in Cape Cod Bay in October.
It is not likely that these wandering salmon return at all to their home rivers; probably they [page 128] are lost permanently from the breeding population. But the much greater numbers that remain localized not very far from their parent streams are believed to follow about the same routes on their return journeys that they followed when they went to sea. Thus, only a few are caught on the Nova Scotia shore between the entrance to St. Mary's Bay and Digby Gut, but fish en route to the Shubenacadie River system are taken in some numbers as they follow the shore of Annapolis and Kings Counties (the Annapolis River also yields a few salmon in its lower course, and some are taken in the Annapolis Basin). Similarly, salmon approaching the St. John River strike the coast about Point Lepreau (about 23 miles to the west) and support an important fishery from there to the mouth of the river.
A question closely bound to the movements of salmon to the sea is: what proportion of them return to spawn in the very rivers in which they were hatched? It seems demonstrated by a variety of evidence, especially by the recapture of tagged fish, that the majority do return. Huntsman, for example, reports an extraordinary instance, of a kelt taken from the Sackville River on the outer coast of Nova Scotia that was tagged and released in the Shubenacadie River system at the head of the Bay of Fundy, and then found its way out of the Bay, around the Nova Scotia coast, and back again to the Sackville, where it was recaptured. We can only speculate how it directed its course, and why it did not turn in to the mouth of any of the other salmon rivers it passed en route. On the other hand, marked fish are sometimes caught in strange rivers. Fish, for instance, that were tagged in Minas Channel have been caught later in the St. John River. And odd fish appear from time to time in rivers where no salmon have been hatched for many years (in the Merrimac for instance).
In short, the parent-stream theory does not always hold. Probably the truth is that while most of the fish never stray far away and do return to the home stream, wanderers that chance, in the spring, to be in the physical state leading to maturity may enter any unpolluted stream they encounter, no matter how far from home.
Dr. Huntsman's studies, carried on through many years, make it increasingly probable that the journeyings of our salmon in salt water are not the result of purposeful swimming in a definite direction, but that they tend to drift with the current as herring do (p. 97), so that the direction in which they travel depends chiefly on the depth at which they happen to be, in relation to the differential circulation of the water at different levels. If so, the St. John River fish tend to drift out with the river water as they scatter. And most of them do appear to remain more or less concentrated in the mid-depths where the principal mixing takes place between the river discharge and the water of the open Bay of Fundy, some 20 to 30 miles from St. John Harbor, living where they find an abundance of herring of various sizes as food. Here Dr. Huntsman calculates the space for them is so great that no two of the approximately 50,000 fish that comprise the total yearly catch need be closer to each other than three-quarters of a mile in a layer of water 5 feet thick; so there is no crowding. But the tagging experiments have shown that the fish that go to sea from Minas Channel, where the outflow is not so definitely localized, scatter more widely, some of them drifting right around the Bay of Fundy with the anti-clockwise circulation.
The situation is not so clear for the coast of Maine, partly because of the paucity of present day information, partly because the several rivers there that once had runs of salmon are so closely spaced along the coast that it is not possible to evaluate their individual contributions to the yearly catches.
With the relationship between salmon journeys and water movements so extremely complex, all we dare say in this regard is that the inshore drift of the deeper layers (characteristic of circulation of the estuarine type) and the slackening of the offshore drift of the fresher surface water that is to be expected as the spring freshets diminish, may be the cause, at least in part, for bringing the salmon into the estuaries, and close inshore elsewhere, in spring. But the nature of the stimulus that impels a salmon to enter fresh water, and then fight his or her way upstream, remains a mystery.
It is not known whether all the salmon move inshore in spring, or only those that are destined [page 129] to spawn that year, plus a certain number of immature grilse that have passed 1 year at sea. And Dr. Huntsman has pointed out that the movement of the salmon riverward may be very slow; thus the salmon may take as much as a month to cover the 20-odd miles to the head of tide in the Petitcodiac River, while some of those that enter the estuary of the St. John River in autumn pass the winter there (probably in a lethargic state) before moving up to the head of tide 80 miles distant. In any ease, only such fish as are approaching sexual maturity (irrespective of age), and some immature female grilse, run far up into the rivers; all the others remain in salt water, or at most they do not run above the head of tide, as has often been remarked.
The majority of the Gulf of Maine salmon become "river-mature" as it is called, long before the spawning season, for while none of them spawn before October, some of them enter fresh water as early as March and April. But the chief runs come later, varying in date, not only from river to river, but from year to year in a given river. In the Penobscot, some fish may enter in March; they are to be expected in the lower reaches after the first week in April; more come in May, perhaps two-thirds in June, with a few fish entering later still. Available information is to the effect that few enter the Narraguagus and Dennys Rivers until well into May, the chief runs there coming in June, with some entering as late as September. We have not been able to obtain definite dates for the spring and early summer runs in the St. John River. But it seems that salmon continue to enter the latter until well into the autumn, judging from catches of fish so fat that they must have come in recently from the sea. Salmon enter other streams tributary to the Bay of Fundy from May on. As a rule the large salmon come earliest, the grilse not until later, probably because it is not until later that the latter have reached the degree of fatness associated with river maturity. Accordingly, the heaviest runs in the Shubenacadie, mostly grilse (p. 130), are said to come from August until late in the autumn.
Every salmon fisherman is familiar with the fact that salmon enter in "runs" that are spaced irregularly in time, and that vary in date from year to year, depending on the height of water in the river and on the strength of the current. Freshets tend to bring them in; if the current becomes too strong they simply hold position, to breast it again as the flow slackens. The fish that are in the estuary remain there during the periods between freshets, waiting, as it were, for the message from upstream that starts them on their way. And the salmon within the river are similarly quiescent during periods of low water and weak current. This is the chief reason why salmon angling is so uncertain a sport, even in the best of rivers.
A good deal of discussion has centered about the question whether the earliest fish stay in fresh water from then until spawning time (a matter of 6 months) or whether there is more or less movement in and out of the river mouths at the beginning of the season. The latter view may be correct for the small streams, but it seems safe to say that after the run is well under way in late May or early June no fish return from fresh to salt water until autumn. Tagging experiments carried out in Canadian rivers have also yielded the very interesting information that no matter when a salmon runs upstream in one year, it may do so either early or late in the next.
It is a matter of common knowledge that salmon average larger in some rivers than in others, and growth studies based on the scales have shown that these differences are due chiefly to the average ages of the fish that enter. In the St. John, as Huntsman has pointed out, there are three principal groups of salmon: (a) male grilse, averaging about 6 pounds, that are mature and fated to breed that same autumn; (b) the ordinary spawners that have passed two years or more at sea; these average 10 to 15 pounds in weight and enter from May to August, the late comers running heavier than the early comers; most of them are virgins, but some of them have already spawned once or twice; (c) immature female grilse, averaging about 9 pounds, that enter from November to January. Few, however, return to spawn in the rivers of Maine until they have passed 2 years at sea; not more than 3 or 4 grilse to 70 adults were taken in the St. Croix, for example, when there still was a good run there, and not more than 1 [page 130] grilse to 500 adults in the Dennys and Penobscot Rivers.
The average weight of the salmon caught in the Penobscot was about 11½ pounds in 1905 (6,378 fish), 9 to 10 pounds in 1919 and 1920 (3,920 fish), or a little less than in the St. John. The heaviest Penobscot fish of which we found definite record of late years weighed a little more than 35 pounds. The fish in the rivers flowing into the head of the Bay of Fundy run much smaller, as Perley pointed out long ago, and recent studies show that most of them spawn first as grilse, i. e., after only one year at sea; a few, having spawned after one year at sea, return to spawn again a year later; and the percentage of larger and older fish is very small there. This, Huntsman points out, contrasts with the prevalent 6-year-old fish in the Miramichi, which discharges into the southern side of the Gulf of St. Lawrence, and with 7- or 8-year-old fish in the Grand Cascapedia, tributary to the Bay of Chaleur. Various explanations have been advanced to account for these differences from river to river, none of them convincing in our opinion.
It also appears to be true (as often stated) that a larger proportion of the salmon are annual spawners in small streams, where most of the spent fish drop downstream again soon after spawning, than in large rivers where many of these "kelts" remain in fresh water over the winter. A plausible explanation is that kelts that return to the sea immediately after spawning have less opportunity to grow (though they recover condition sufficiently to spawn again the following summer) than such as await the spring to go downstream, and that spend a whole year at sea instead of one winter only between two succesive spawnings. This, however, does not account for the fact that it is almost invariably the large rivers that yield the very large maiden fish that have spent 4 years at sea, or more.
The early extirpation of salmon from the Merrimac, Saco, Kennebec system, and various rivers to the eastward naturally resulted in a great decrease in the abundance of salmon in the open Gulf, clearly reflected in the catches. Data are not available for early years when all the rivers still offered free access. But the yearly catch had been reduced to about 100 to 500 fish in the St. Croix by about 1887; 200 to 1,000 each in the Dennys and in the Kennebec, and 5,000 to 15,000 in the Penobscot. The catch along the Maine coast, which had been a little more than 150,000 pounds in 1889 (more than seven-eighths of this in or about the approaches to the Penobscot), was only about 86,000 pounds in 1905 (of this 74,000 lb., or 6,378 fish from the Penobscot); was about 20,000 pounds in 1919 (13,557 lb. or 1,322 fish from the Penobscot); and was only 14,744 pounds (12,700 lb. or 1,221 fish from the Penobscot) in 1928. As 70 to 90 percent of the Maine catch comes from Penobscot River or Bay, the following table of salmon caught there in certain years from 1896 to 1928 is pertinent:
|year||number of fish||pounds|
The Maine catch then increased again to about 88,000 pounds in 1930 and to about 70,000 in 1931, suggesting a better run in the Penobscot, and varied between 16,000 and 40,000 pounds through the period 1932-1938. But the average reported catch for Maine for the period 1939 to 1947 was only about 3,600 pounds (maximum 9,300, minimum 600), the average Massachusetts catch for the same period only about 100 pounds (maximum 400, minimum 0). Thus the output of salmon from the rivers of Maine (none from the rivers of Massachusetts) has been only about one-fiftieth as great during the past few years as it was some 60 years ago.
The numbers of salmon have held up much better in the Canadian waters of the Gulf, thanks to wise measures of conservation such as limiting netting at the mouths of the rivers, and keeping the streams free for access by fishways at the dams. The average yearly catches, from 1870 to 1946, [page 131] were as follows for the west coast of Nova Scotia and for the Bay of Fundy combined:
The Canadian catch in the open Gulf and in the Bay of Fundy may be expected to run about 400,000 to 600,000 pounds at the present time, taking one year with another, or something like 40,000 to 60,000 fish, which is perhaps 100 times as great as that for the entire coastline of Maine and of Massachusetts. And the distribution of the catches shows that the St. John River contributes something like four-fifths of this, or a yearly average of some 50,000 fish, contrasting with only a few hundred fish for the Penobscot in a poor year, and perhaps up to 8,000 in a good.
Salmon anglers are only too familiar with the fact that the number of fish that enter even the best of salmon rivers is much smaller in some years than in others. During the 16-year period, 1931-1946, the commercial catches reported for St. John Harbor and St. John River (best salmon river tributary to the Gulf of Maine) were good in 1931 (164,000 lbs.); in 1935 (149,300 lbs.); in 1936 (148,600 lbs.); in 1937 (172,700 lbs.); and in 1943 (157,500 lbs.); but were poor in 1939 (48,500 lbs.); in 1945 (60,000 lbs.) and in 1946 (54,500 lbs.). The yearly average for this period was 116,000 pounds.
In the Minas system the fishery produced as much as 383,800 pounds in 1907, 283,400 pounds in 1917, and 226,500 pounds in 1918; but since then, up to 1946, the best catches have been only 160,700 pounds in 1919, 165,100 pounds in 1923, and 143,300 pounds in 1925, while the poorest were 28,100 pounds in 1938 and 26,600 pounds in 1945. The average yearly catch from 1917 to 1930 was 133,000 pounds, and from 1931 to 1946, 48,000 pounds.
The reader will notice at once that the big years have not been the same for these two bodies of salmon. It seems sufficiently established that yearly and regional differences, such as these, result in the main from corresponding differences in the numbers of smolts, that reach salt water in any given year. And recent investigations in Canadian waters make it likely that the factor chiefly responsible is the height of the water from summer to summer, or over periods of several summers, which of course reflects the yearly variations in rainfall. If the water is high the parr are protected from the birds that prey upon them and are more easily able to escape the trout, so that many survive to descend to the sea and to return one, two, or three years later. If the water in the river is low the parr are more at the mercy of kingfishers, megansers, and trout, so that fewer of them live to reach salt water, and there are fewer of them to return as grilse or as older fish.
 Huntsman (Bull. Biol. Board Canada, 21, 1931) has published an extensive study of the life history of the salmon of the Maritime Provinces of Canada, from which we have drawn freely in the following account. See also Huntsman and others (Migration and Conserv. of Salmon, Pub. No. 8, Amer. Assoc. Adv. Sci., 1939) for discussions of the movements of the salmon in Canadian and Newfoundland waters; also Lindsay and Thompson (Rept. Newfoundland Fish. Res. Comm., vol. 1, No. 2, 1932) for an account of the biology of the salmon in the rivers and around the coasts of Newfoundland.
 Trans. Amer. Fish. Soc., vol. 24, 1934, p. 211.
 Mem. Boston Soc. Nat. Hist., vol. 9, No. 1, 1935, p. 65-68.
 They are voracious now, and fly-fishing for these "black salmon" as they are called, is a favorite sport nowadays, especially in Miramichi waters tributary to the Gulf of St. Lawrence.
 Huntsman, Bull 21, Biol. Bd. Canada, 1931, p. 81. based on studies by Kerr and by Blair.
 Life and Sport on the North Shore, 1909, Quebec.
 Kendall (Mem. Boston Soc. Nat. Hist., vol. 9, No. 1, 1935, p. 34) found smelts in Penobscot salmon, alewives in salmon from the St. John.
 See Kendall (Mem. Boston Soc. Nat. Hist., vol. 9, No. 1, 1935, pp. 33-34) for a recent survey of the diet of salmon in general; the Gulf of Maine fish in particular, with references. Eichelbaum (Cons. Perm. Internat. Explor. Mer, Rapports et Proc. Verb., vol. 21, 1916, p. 84) records the contents of many salmon from the Baltic and from the North Sea.
 Huntsman (Bull. Biol. Board Canada, No. 21, 1931) gives an interesting account of these tagging experiments, from which this summary is drawn.
 As worked out from its scales by Dr. D. L. Belding, and reported in Field and Stream, August 1951, p. 10.
 It is commonly stated that this applies chiefly to the males. But Huntsman (Bull. Biol. Board Canada, No. 21, 1931, pp. 18-19) has found that grilse of both sexes spawn in the small rivers at the head of the Bay of Fundy.
 Kendall. Mem. Boston Soc. Nat. Hist., vol. 9, No. 1, 1935, p. 32.
 This is the southern European limit given by Berg (Zoogeographica. vol. 1, Pt. 2, 1932, p. 112.)
 Jensen, Fauna of Greenland, vol. 1, Pt. 3, Fishes, 1928, pp. 3 and 4, Copenhagen.
 Vladykov (Contrib. Canad. Biol., N. Ser., vol. 8, No. 2, 1933, p. 18, fig. 1) shows a locality record near Fort Chimo, and there are salmon in the rivers of the eastern part of Ungava Bay.
 Blair (Res, Bull. 12, Dept. Nat. Resources Newfoundland, 1943, pp. 5-17) gives a detailed account of the salmon rivers of the outer Labrador coast, Strait of Belle Isle to Hamilton Inlet.
 Atkins (1887, Fish. Ind. U. S., Sect. 5, vol. 1, p. 679) has collected much information on the local history of salmon in northern New England.
 Lyman and Reed, Rept. Comm. Fish. Massachusetts (1865) 1866, Senate Doc. 8, pp. 36-41.
 A few small "salmon" reported of late in the Merrimac probably were the landlocked form, running down from tributaries stocked with this fish.
 Three reports of salmon caught on Western Bank have appeared in the daily press since 1925 to our knowledge, and Kendall (Mem. Boston Soc. Nat. Hist., vol. 9, No. 1, 1935, p. 33) reports one caught on La Have Bank 100 miles from Halifax, and another 60 miles off Cape Sable.
 Kendall (Mem. Boston Soc. Nat. Hist., vol. 9, No. 1, 1935, pp. 31-33) lists a number of such cases.
 Huntsman, Ann. Rept. Fish. Res. Board Canada, (1947) 1948, p. 37, and unpublished notes.
 The average was only 3,000 pounds (perhaps 300 fish) for the years 1939, 1940, 1943, 1944. Statistics are not readily available for 1941 and 1942.
 1933, 1935, 1937, 1938, 1939, 1940, 1943, 1944, 1945, and 1946.
 Information contributed by the Pond Village Cold Storage Co.
 In the spring of 1915 about 75 (including fish up to 35 lb.) were taken at Gay Head and in the neighborhood of Woods Hole.
 Smith (Bull. U. S. Fish. Comm., vol. 14, 1895, p. 99) reports salmon seined among some mackerel off Delaware in 1893.
 Bull. 51, Biol. Board of Canada, 1936, p. 9.
 See Huntsman, Pub. Amer. Assoc. Adv. Sci. 8, 1932, p. 35, for summary of these records.
 Huntsman, Science, vol. 95, 1947, p. 381.
 Huntsman, Ann. Rept. Fish. Res. Bd. Canada, (1947) 1948, p. 37.
 Huntsman, Science, vol. 85, 1937, p. 314; Pub. 8, Amer. Assoc. Adv. Sci., 1939, p. 35.
 Ann. Rept. Fish. Res. Bd. Canada (1947) 1948, p. 38.
 Huntsman, Ann. Rept. Fish. Res. Bd. Canada (1948) 1949, p. 40.
 Bulletin 21, Biol. Bd. Canada, 1931, p. 96.
 This was shown by Huntsman, Ann. Rep. Fish. Res. Bd. Canada (1947) 1948, p. 37.
 Progress Report, Atlantic stations, Biol. Bd. Canada, 8, 1933, p. 6; and unpublished notes.
 Fifty-fifth Annual Report of the Fisheries Branch, Department of Marine and Fisheries, Canada, (1921-22) 1922, p. 19.
 Nature, vol. 141, 1938, p. 421; Pub. 8, Amer. Assoc. Adv. Sci., 1939, p. 34.
 See Kendall (Mem. Boston Soc. Nat. Hist., vol. 9, No. 1, 1935, pp. 58-60) for age determinations of Penobscot salmon.
 Radcliffe, Rept. U. S. Comm. Fish. (1921) 1922, p. 146.
 Kendall, Mem. Boston Soc. Nat. Hist., vol. 9. No. 1, 1935, p. 32.
 Bull. 21, Biol. Board Canada, 1931. p. 19.
 No data are available for 1934 or 1936.
 No data for 1941.
 Huntsman (Bull. 21, Biol. Board Canada, 1931) has made a very interesting analysis of catches for the Bay of Fundy as a whole, as well as for the St. John, for the Chignecto system, and for the Minas system, separately.