Silent Spring

by Rachel Carson

  So in 1954, in the month of June, the planes visited the forests of the Northwest Miramichi and white clouds of settling mist marked the crisscross pattern of their flight. The spray—one-half pound of DDT to the acre in a solution of oil—filtered down through the balsam forests and some of it finally reached the ground and the flowing streams. The pilots, their thoughts only on their assigned task, made no effort to avoid the streams or to shut off the spray nozzles while flying over them; but because spray drifts so far in even the slightest stirrings of air, perhaps the result would have been little different if they had.

  Soon after the spraying had ended there were unmistakable signs that all was not well. Within two days dead and dying fish, including many young salmon, were found along the banks of the stream. Brook trout also appeared among the dead fish, and along the roads and in the woods birds were dying. All the life of the stream was stilled. Before the spraying there had been a rich assortment of the water life that forms the food of salmon and trout—caddis fly larvae, living in loosely fitting protective cases of leaves, stems or gravel cemented together with saliva, stonefly nymphs clinging to rocks in the swirling currents, and the wormlike larvae of blackflies edging the stones under riffles or where the stream spills over steeply slanting rocks. But now the stream insects were dead, killed by the DDT, and there was nothing for a young salmon to eat.

  Amid such a picture of death and destruction, the young salmon themselves could hardly have been expected to escape, and they did not. By August not one of the young salmon that had emerged from the gravel beds that spring remained. A whole year's spawning had come to nothing. The older young, those hatched a year or more earlier, fared only slightly better. For every six young of the 1953 hatch that had foraged in the stream as the planes approached, only one remained. Young salmon of the 1952 hatch, almost ready to go to sea, lost a third of their numbers.

  All these facts are known because the Fisheries Research Board of Canada had been conducting a salmon study on the Northwest Miramichi since 1950. Each year it had made a census of the fish living in this stream. The records of the biologists covered the number of adult salmon ascending to spawn, the number of young of each age group present in the stream, and the normal population not only of salmon but of other species of fish inhabiting the stream. With this complete record of prespraying conditions, it was possible to measure the damage done by the spraying with an accuracy that has seldom been matched elsewhere.

  The survey showed more than the loss of young fish; it revealed a serious change in the streams themselves. Repeated sprayings have now completely altered the stream environment, and the aquatic insects that are the food of salmon and trout have been killed. A great deal of time is required, even after a single spraying, for most of these insects to build up sufficient numbers to support a normal salmon population—time measured in years rather than months.

  The smaller species, such as midges and blackflies, become re-established rather quickly. These are suitable food for the smallest salmon, the fry only a few months old. But there is no such rapid recovery of the larger aquatic insects, on which salmon in their second and third years depend. These are the larval stages of caddis flies, stoneflies, and mayflies. Even in the second year after DDT enters a stream, a foraging salmon parr would have trouble finding anything more than an occasional small stonefly. There would be no large stoneflies, no mayflies, no caddis flies. In an effort to supply this natural food, the Canadians have attempted to transplant caddis fly larvae and other insects to the barren reaches of the Miramichi. But of course such transplants would be wiped out by any repeated spraying.

  The budworm populations, instead of dwindling as expected, have proved refractory, and from 1955 to 1957 spraying was repeated in various parts of New Brunswick and Quebec, some places being sprayed as many as three times. By 1957, nearly 15 million acres had been sprayed. Although spraying was then tentatively suspended, a sudden resurgence of budworms led to its resumption in 1960 and 1961. Indeed there is no evidence anywhere that chemical spraying for budworm control is more than a stopgap measure (aimed at saving the trees from death through defoliation over several successive years), and so its unfortunate side effects will continue to be felt as spraying is continued. In an effort to minimize the destruction of fish, the Canadian forestry officials have reduced the concentration of DDT from the ½ pound previously used to ¼ pound to the acre, on the recommendation of the Fisheries Research Board. (In the United States the standard and highly lethal pound-to-the-acre still prevails.) Now, after several years in which to observe the effects of spraying, the Canadians find a mixed situation, but one that affords very little comfort to devotees of salmon fishing, provided spraying is continued.

  A very unusual combination of circumstances has so far saved the runs of the Northwest Miramichi from the destruction that was anticipated—a constellation of happenings that might not occur again in a century. It is important to understand what has happened there, and the reasons for it.

  In 1954, as we have seen, the watershed of this branch of the Miramichi was heavily sprayed. Thereafter, except for a narrow band sprayed in 1956, the whole upper watershed of this branch was excluded from the spraying program. In the fall of 1954 a tropical storm played its part in the fortunes of the Miramichi salmon. Hurricane Edna, a violent storm to the very end of its northward path, brought torrential rains to the New England and Canadian coasts. The resulting freshets carried streams of fresh water far out to sea and drew in unusual numbers of salmon. As a result, the gravel beds of the streams which the salmon seek out for spawning received an unusual abundance of eggs. The young salmon hatching in the Northwest Miramichi in the spring of 1955 found circumstances practically ideal for their survival. While the DDT had killed off all stream insects the year before, the smallest of the insects—the midges and blackflies—had returned in numbers. These are the normal food of baby salmon. The salmon fry of that year not only found abundant food but they had few competitors for it. This was because of the grim fact that the older young salmon had been killed off by the spraying in 1954. Accordingly, the fry of 1955 grew very fast and survived in exceptional numbers. They completed their stream growth rapidly and went to sea early. Many of them returned in 1959 to give large runs of grilse to the native stream.

  If the runs in the Northwest Miramichi are still in relatively good condition this is because spraying was done in one year only. The results of repeated spraying are clearly seen in other streams of the watershed, where alarming declines in the salmon populations are occurring.

  In all sprayed streams, young salmon of every size are scarce. The youngest are often "practically wiped out," the biologists report. In the main Southwest Miramichi, which was sprayed in 1956 and 1957, the 1959 catch was the lowest in a decade. Fishermen remarked on the extreme scarcity of grilse—the youngest group of returning fish. At the sampling trap in the estuary of the Miramichi the count of grilse was only a fourth as large in 1959 as the year before. In 1959 the whole Miramichi watershed produced only about 600,000 smolt (young salmon descending to the sea). This was less than a third of the runs of the three preceding years.

  Against such a background, the future of the salmon fisheries in New Brunswick may well depend on finding a substitute for drenching forests with DDT.

  The eastern Canadian situation is not unique, except perhaps in the extent of forest spraying and the wealth of facts that have been collected. Maine, too, has its forests of spruce and balsam, and its problem of controlling forest insects. Maine, too, has its salmon runs—a remnant of the magnificent runs of former days, but a remnant hard won by the work of biologists and conservationists to save some habitat for salmon in streams burdened with industrial pollution and choked with logs. Although spraying has been tried as a weapon against the ubiquitous budworm, the areas affected have been relatively small and have not, as yet, included important spawning streams for salmon. But what happened to stream fish in an area observed by the Maine Depa
rtment of Inland Fisheries and Game is perhaps a portent of things to come.

  "Immediately after the 1958 spraying," the Department reported, "moribund suckers were observed in large numbers in Big Goddard Brook. These fish exhibited the typical symptoms of DDT poisoning; they swam erratically, gasped at the surface, and exhibited tremors and spasms. In the first five days after spraying, 668 dead suckers were collected from two blocking nets. Minnows and suckers were also killed in large numbers in Little Goddard, Carry, Alder, and Blake Brooks. Fish were often seen floating passively downstream in a weakened and moribund condition. In several instances, blind and dying trout were found floating passively downstream more than a week after spraying."

  (The fact that DDT may cause blindness in fish is confirmed by various studies. A Canadian biologist who observed spraying on northern Vancouver Island in 1957 reported that cutthroat trout fingerlings could be picked out of the streams by hand, for the fish were moving sluggishly and made no attempt to escape. On examination, they were found to have an opaque white film covering the eye, indicating that vision had been impaired or destroyed. Laboratory studies by the Canadian Department of Fisheries showed that almost all fish [Coho salmon] not actually killed by exposure to low concentrations of DDT [3 parts per million] showed symptoms of blindness, with marked opacity of the lens.)

  Wherever there are great forests, modern methods of insect control threaten the fishes inhabiting the streams in the shelter of the trees. One of the best-known examples of fish destruction in the United States took place in 1955, as a result of spraying in and near Yellowstone National Park. By the fall of that year, so many dead fish had been found in the Yellowstone River that sportsmen and Montana fish-and-game administrators became alarmed. About 90 miles of the river were affected. In one 300-yard length of shoreline, 600 dead fish were counted, including brown trout, whitefish, and suckers. Stream insects, the natural food of trout, had disappeared.

  Forest Service officials declared they had acted on advice that 1 pound of DDT to the acre was "safe." But the results of the spraying should have been enough to convince anyone that the advice had been far from sound. A cooperative study was begun in 1956 by the Montana Fish and Game Department and two federal agencies, the Fish and Wildlife Service and the Forest Service. Spraying in Montana that year covered 900,000 acres; 800,000 acres were also treated in 1957. The biologists therefore had no trouble finding areas for their study.

  Always, the pattern of death assumed a characteristic shape: the smell of DDT over the forests, an oil film on the water surface, dead trout along the shoreline. All fish analyzed, whether taken alive or dead, had stored DDT in their tissues. As in eastern Canada, one of the most serious effects of spraying was the severe reduction of food organisms. On many study areas aquatic insects and other stream-bottom fauna were reduced to a tenth of their normal populations. Once destroyed, populations of these insects, so essential to the survival of trout, take a long time to rebuild. Even by the end of the second summer after spraying, only meager quantities of aquatic insects had re-established themselves, and on one stream—formerly rich in bottom fauna—scarcely any could be found. In this particular stream, game fish had been reduced by 80 per cent.

  The fish do not necessarily die immediately. In fact, delayed mortality may be more extensive than the immediate kill and, as the Montana biologists discovered, it may go unreported because it occurs after the fishing season. Many deaths occurred in the study streams among autumn spawning fish, including brown trout, brook trout, and whitefish. This is not surprising, because in time of physiological stress the organism, be it fish or man, draws on stored fat for energy. This exposes it to the full lethal effect of the DDT stored in the tissues.

  It was therefore more than clear that spraying at the rate of a pound of DDT to the acre posed a serious threat to the fishes in forest streams. Moreover, control of the budworm had not been achieved and many areas were scheduled for respraying. The Montana Fish and Game Department registered strong opposition to further spraying, saying it was "not willing to compromise the sport fishery resource for programs of questionable necessity and doubtful success." The Department declared, however, that it would continue to cooperate with the Forest Service "in determining ways to minimize adverse effects."

  But can such cooperation actually succeed in saving the fish? An experience in British Columbia speaks volumes on this point. There an outbreak of the black-headed budworm had been raging for several years. Forestry officials, fearing that another season's defoliation might result in severe loss of trees, decided to carry out control operations in 1957. There were many consultations with the Game Department, whose officials were concerned about the salmon runs. The Forest Biology Division agreed to modify the spraying program in every possible way short of destroying its effectiveness, in order to reduce risks to the fish.

  Despite these precautions, and despite the fact that a sincere effort was apparently made, in at least four major streams almost 100 per cent of the salmon were killed.

  In one of the rivers, the young of a run of 40,000 adult Coho salmon were almost completely annihilated. So were the young stages of several thousand steelhead trout and other species of trout. The Coho salmon has a three-year life cycle and the runs are composed almost entirely of fish of a single age group. Like other species of salmon, the Coho has a strong homing instinct, returning to its natal stream. There will be no repopulation from other streams. This means, then, that every third year the run of salmon into this river will be almost nonexistent, until such time as careful management, by artificial propagation or other means, has been able to rebuild this commercially important run.

  There are ways to solve this problem—to preserve the forests and to save the fishes, too. To assume that we must resign ourselves to turning our waterways into rivers of death is to follow the counsel of despair and defeatism. We must make wider use of alternative methods that are now known, and we must devote our ingenuity and resources to developing others. There are cases on record where natural parasitism has kept the budworm under control more effectively than spraying. Such natural control needs to be utilized to the fullest extent. There are possibilities of using less toxic sprays or, better still, of introducing microorganisms that will cause disease among the budworms without affecting the whole web of forest life. We shall see later what some of these alternative methods are and what they promise. Meanwhile, it is important to realize that chemical spraying of forest insects is neither the only way nor the best way.

  The pesticide threat to fishes may be divided into three parts. One, as we have seen, relates to the fishes of running streams in northern forests and to the single problem of forest spraying. It is confined almost entirely to the effects of DDT. Another is vast, sprawling, and diffuse, for it concerns the many different kinds of fishes—bass, sunfish, crappies, suckers, and others—that inhabit many kinds of waters, still or flowing, in many parts of the country. It also concerns almost the whole gamut of insecticides now in agricultural use, although a few principal offenders like endrin, toxaphene, dieldrin, and heptachlor can easily be picked out. Still another problem must now be considered largely in terms of what we may logically suppose will happen in the future, because the studies that will disclose the facts are only beginning to be made. This has to do with the fishes of salt marshes, bays, and estuaries.

  It was inevitable that serious destruction of fishes would follow the widespread use of the new organic pesticides. Fishes are almost fantastically sensitive to the chlorinated hydrocarbons that make up the bulk of modern insecticides. And when millions of tons of poisonous chemicals are applied to the surface of the land, it is inevitable that some of them will find their way into the ceaseless cycle of waters moving between land and sea.

  Reports of fish kills, some of disastrous proportions, have now become so common that the United States Public Health Service has set up an office to collect such reports from the states as an index of water pollution.

/>   This is a problem that concerns a great many people. Some 25 million Americans look to fishing as a major source of recreation and another 15 million are at least casual anglers. These people spend three billion dollars annually for licenses, tackle, boats, camping equipment, gasoline, and lodgings. Anything that deprives them of their sport will also reach out and affect a large number of economic interests. The commercial fisheries represent such an interest, and even more importantly, an essential source of food. Inland and coastal fisheries (excluding the offshore catch) yield an estimated three billion pounds a year. Yet, as we shall see, the invasion of streams, ponds, rivers, and bays by pesticides is now a threat to both recreational and commercial fishing.

  Examples of the destruction of fish by agricultural crop sprayings and dustings are everywhere to be found. In California, for example, the loss of some 60,000 game fish, mostly bluegill and other sunfish, followed an attempt to control the rice-leaf miner with dieldrin. In Louisiana 30 or more instances of heavy fish mortality occurred in one year alone (1960) because of the use of endrin in the sugarcane fields. In Pennsylvania fish have been killed in numbers by endrin, used in orchards to combat mice. The use of chlordane for grasshopper control on the high western plains has been followed by the death of many stream fish.

  Probably no other agricultural program has been carried out on so large a scale as the dusting and spraying of millions of acres of land in southern United States to control the fire ant. Heptachlor, the chemical chiefly used, is only slightly less toxic to fish than DDT. Dieldrin, another fire ant poison, has a well-documented history of extreme hazard to all aquatic life. Only endrin and toxaphene represent a greater danger to fish.