parts of the stream their condition remains much lower than that of the small permanent residents of the river. The remarkable thing about the ill-conditioned trout of 1932 is not that they failed to regain condition in the river, but that they remained upstream so long after the spawning season when there was no obstacle to their downward passage, and it was for the purpose of determining the cause of this abnormal behaviour that specimens were collected and examined.

Upon dissection the group proved to be composed of eighteen females and two males, the latter being among the specimens taken upstream. The presence of fully-developed eggs, ranging in number from nine to one hundred and forty-six, in all females examined indicated that all had reached maturity and that most of the eggs developed by each fish had been parted with. The majority of the eggs present were lying free in the body cavity, but in some specimens a few eggs were retained within the ovaries; others had eggs lodged between the liver and the structure that is referred to by Gunther (1880) as the diaphrama, by Kendall (1920) as the diaphram, and by Percival (loc. cit., p. 346) as the abdomino-pericardial wall, and in four there were eggs embedded in the liver. The latter organ was, in one specimen, much broken up, with eggs present in the interstices of the mass. More than half the specimens examined showed discolouration of parts of the liver, usually associated with a roughening of the surface and a shrinkage of the parts affected. These parts when sectioned, revealed considerable breaking down of structure. The ovaries of all females examined were damaged in a greater or less degree. In one specimen the left mesovarium was parted from its anterior and dorsal anchorage for about three-quarters of its length and was turned back with its anterior end to the rear of the body cavity, the partly detached membrane showing several rents and openings. The right mesovarium, which had retained its position, had five rents extending transversely from the free edge, two of these rents having formed a flap which had turned inwards and appeared to be in the process of forming an egg-retaining pocket such as is frequently observed in Lake Ellesmere trout. The least damaged specimen had one ovary uninjured and only a slight transverse rent toward the rear of the other. Between this extreme and the specimen first described there was complete intergradation, the injuries being identical in character and differing only in severity. The genital organs of the two males examined showed no definite evidence of injury.

In three specimens the lower intestinal mesentery disclosed injuries consisting of holes through the membrane, partings from the point of anchorage and the tearing back of the vertical face, and in two the upper intestinal mesentery was injured in a similar manner. These injuries were usually almost healed and had obviously been inflicted some time previously. Eight specimens revealed one or more ruptures of the peritoneum, most of which were in process of healing, but in two the structure of a considerable area of peritoneum together with that of much of the adjacent muscle was completely broken down. The continuance in life of these fish must necessarily have been brief.

In most specimens the muscular structure of the abdominal wall showed considerable inflammation, this being most severe in the region between the ventral fins and the genital outlet. Three specimens, including the two males, showed only slight inflammation, which would probably have passed away in a short time. One specimen had a large portion of the spleen almost severed from the remainder while in two others small portions had been detached at the porterior extremity. In a few instances the intestinal wall showed breaking down of structure.

Externally the fish showed no trace of recent injury. Two specimens possessed old scars, but the scale covering had been completely replaced on the parts affected, and by examination of the replacement scales the date of injury was determined to be over two years previously in one specimen and nearly two years in the other, the date of injury in each instance preceding the attainment of maturity. A similar percentage of scarred fish is normal in all collections of mature Lake Ellesmere trout. With the exception of one specimen taken near the river mouth, which contained seven smelts, the stomachs of the fish examined were almosh empty.

The ill-conditioned fish remained upstream in the pools and continued to die off until October 25 when a flood occurred and they were washed away, their places being taken later by stragglers from the Hororata and the upper permanent water of the main stream. No further flood occurred during the summer, autumn and winter, and in the spring of 1933 conditions were indentical with those of the previous year. It is stated on page 348 of the paper under discussion that in September, 1933, there were many large ill-conditioned fish lying in pools about ten miles from the river mouth, and that these fish would not leave the pool although there was free egress; and it is further stated that at the end of Sepember a slight freshet occurred and the fish disappeared. The latter statement is not correct. There was no freshet, rise or discolouration of the water in September, 1933, and there was no noticeable movement of fish in this month or the following month. At the opening of the fishing season on October 1st the river was very low with an abnormal development of algae, and the pools in the locality indicated contained a greater number of large ill-conditioned fish than had been present in the previous year. Reference to the writer's diary shows that on October 1st, 1933, over thirty large fish were visible in Butterfield's Pool, about twenty in Plunket's Pool, nineteen in the Meadowbank Pool, and over seventy in the Clay Block Pool. This pool affords considerable facilities for the concealment of fish, and it is probable that double the number counted was present. The Aquarium carried a large shoal, the full extent of which could not be determined on account of the deep shade, there were eleven fish in the Moonlight Pool, about twenty-five in the Swirl, between thirty and forty in McGregor's Pool and a small shoal in the Secrecy Pool. The intervening sections of the river carried many individual specimens and small groups. The fish gradually dwindled away during the first two months of the season, some dying and many being taken by anglers. Specimens were procured by the writer on the 2nd and 5th of October and, on examination, were found to be identical with

A further suggestion made on page 348 of the paper under discussion, that injury may have been suggested by the somewhat ragged appearance of the ovaries, which, it is stated, normally remain flabby for many weeks after the eggs are shed, is irreconcilable with the statement made on page 345 that shrinkage of the ovary occurs before the eggs are discharged from the body, and cannot be considered.

In addition to these suggested explanations of the injuries present in the ill-conditioned trout of 1932, evidence is presented in opposition to the present writer's conclusions that stripping was responsible. It is implied on pages 347 and 350 of the paper under discussion that trout do not proceed upstream after being stripped, and that injuries present in fish collected some miles above the trap could not have been caused by stripping.

Discussing the artificial spawning of trout, Mr W. H. Armistead, proprietor of the Solway fishery, and one of the leading fish-culturists in Great Britain, states (1908):—

Spawned fish should be put into the stream above the trap, if there is no fear of poachers. Most of them will journey up to the spawning beds and deliberately go through the whole performance of making redds, etc., as though they were quite unaware of the fact that their eggs had been taken from them. Speaking on the same subject during the discussion on Kendall's (1920) anatomical paper, Mr J. W. Titcombe, fish culturist, Conservation Commission, Albany, New York, says:—

In the case of wild fish held in pens and stripped from day to day, I have found that if in stripping we leave one or two eggs the trout will stay around the spawning bed until it gets rid of those eggs. We find they are very persistent. We have penned fish two miles from the spawning beds and taken what we believed to be all their eggs, and within twenty-four hours we have found those same fish over on the spawning beds two miles away. I inferred that these fish were there to get rid of the two or three eggs we had left behind in the stripping process.

This statement is referred to by Percival (loc. cit., p. 347) as follows:—

It is probable that Titcombe's phrase “one or two eggs” means perhaps half stripped or thereabouts.

As the grounds upon which this assumption is based are not stated it is unnecessary to attempt any reconciliation between it and Titcombe's high standing as a fish-culturist.^*

A section of the paper under discussion is devoted to a consideration of the ovarian anatomy of trout, and the conclusion is reached therefrom that “it is not clear how eggs could be forced forward in the body cavity when a ripe fish is held upsidedown preparatory to stripping.” As no such suggestion was made by the present writer, the words used being “head downward” (Stokell, 1936, p. 81), no further comment is required.

The suggestion is made (Percival, loc. cit., p. 350) that since the act of stripping females may be regarded as more vigorous than with males, there should be shown some reflection on the sex ratio of fish handed in successive years if stripping were injurious Without presuming to express an opinion on the first contention the present writer submits that, if it is a valid one, a test could more logically be applied to the fish in which injuries existed than to those taken each year in the traps, many of which are necessarily first spawners and cannot have come under the influence suggested. Actually the percentage of females in the injured fish examined, including those from the lake, was 94.

It is further suggested on pages 349 and 351 of the paper under discussion that if stripping were injurious to trout the loss of a number of mature fish would tend to affect the higher age groups, thus causing a reduction in the average size of fish handled at the traps in successive years, and it is contended that as no appreciable reduction in average size is revealed by the table of weights and lengths presented on page 349 the population has remained constant. It is not clear how the extent of a trout population may be judged in this manner when only an inconstant fraction of the spawning run is dealt with and the ages of the fish are unknown. A reduction in the population could be expected to render an improved food supply available to the remaining fish, and this could conceivably be reflected in an improved growth rate and consequently a higher average size. There are also other influences operating at Lake Ellesmere which tend to disqualify any inference of a numerically constant population that may be drawn from a constant average size. The standard mesh of the flounder nets, which permits the passage of small mature trout but retains and kills larger individuals, must be considered as an influence in regulating the average size of adult trout, and the same may be said of the effects of parasitism, which have been found to become progressively more potent in the higher age groups. Quite apart from these considerations the table of average sizes presented in the paper under discussion invalidates itself as a means of disclosing serious loss of mature fish. As it is certain that in 1932 and 1933 fish of the size dealt with in the table died in considerable numbers it is equally certain from the failure of the table to reveal their loss that this method of detection is unsound.

With regard to the contention that egg production has remained constant in successive years the figures given on page 349 of the paper under discussion are so obviously questionable that it would be unsafe to draw any inference from them. If the year 1932 is considered, which is the only one for which particulars of both egg yield and size of fish are given, it will be found that trout stated to average 2.9 pounds in weight are credited with an average production of only 1119 eggs, which works out at approximately 385 eggs per pound of fish. In April, 1937, the present writer counted the eggs of a small but representative group of Lake Ellesmere trout which, though taken at random, had an average weight almost identical with that of the fish just referred to. The average yield of eggs per pound of fish proved low for brown trout, and the individual yields showed considerable differences which, however, appear to be attributable to differences in the sex history of the fish. The most productive specimen was approaching its first maturity while the specimen showing the lowest comparative yield had matured twice previously and revealed evidence of past ovarian injury of the kind described in the second section of this paper. It would thus appear that the low average egg yield of Lake Ellesmere trout is contributed to by the development in individuals of partial sterility consequent on injury by stripping.

The data obtained from the group are given below.

• Average number of eggs per lb. 751.

• Maximum individual number per lb. 960.

• Minimum individual number per lb. 587.

It will thus be seen that the average yield of eggs per fish is 2200, or nearly double the number (1119) recorded by Percival (loc. cit., p. 349) as being obtained by stripping fish of approximately similar size, and that the yield per lb. of fish is 751 as compared with 385, an almost identical ratio. As the figures submitted in the paper under discussion deal with only about half of the eggs produced by Lake Ellesmere trout as they now exist, and as no attempt has been made to ascertain what becomes of the balance, they must be regarded as fractions of unknown constancy from which no deductions can be made. The most probable explanation of the difference between the egg counts and the stripping figures is that many fish

are imperfectly stripped, and, being liberated below the trap, they re-enter and are again stripped and recorded. If this is so it would appear, from the retention of so large a number of eggs as to render the fish indistinguishable as having been stripped, that, in the first instance, stripping is performed before ripeness is complete.

It is now necessary to consider the positive evidence leading up to the conclusion that the injuries were caused by stripping. The absence of injuries from immature Lake Ellesmere fish examined for comparison and from mature river-dwelling fish taken in the vicinity of the pools from which many of the injured fish were taken restricts the injuries to the class of fish handled at the traps, namely, mature lake fish. The presence of stripped fish in pools several miles above the point where the stripping trap is operated is found to be consistent with the behaviour of trout after being subjected to the unnatural process of stripping. The next consideration is that the injuries present in the specimens examined were restricted to the part of the body that is operated upon by strippers. In stripping, the lower part of the body is squeezed from the pectoral fins to the vent, that is to say, the liver, the reproductive organs, the greater part of the alimentary tract, the spleen, the mesenteries, the peritoneum and the muscular structure of the flanks are subjected to pressure. With the exception of the anterior portion of the alimentary tract these are the very structures in which injury existed. The absence of any external injuries such as scars or noticeable derangement of scales disqualifies any explanations based on the grounds of poaching or attacks by predators. Contact with poaching implements, such as spears, hoopiron or snares, if sufficiently severe to cause serious internal injuries, must leave obvious evidence in the form of external marks, and the same applies to attacks by predatory animals, such as shags and large eels. In explaining the association of internal injuries with the absence of external injury it is necessary to postulate a causative agent if a non-rigid character yet capable of exerting considerable pressure—a specification with which human hands completely agree. The further circumstances of time, place and class of fish to which the injuries were restricted leave no alternative to the conclusion that stripping was responsible, and the matter is placed beyond all question by the reproduction of the whole of the conditions described, execept those that are of a secondary nature, in the manner suggested. The simple test of stripping trout with various degrees of pressure and immediately dissecting them showed that serious internal injury could be inflicted without leaving an outward mark. It was unusual to dislodge more than one or two scales. The organs that were found to be the most easily damaged are the ovaries and the liver; then come the mesenteries and the peritoneum. The spleen proved less easily damaged, by reason of its being protected to some extent by the pelvic bones,^* but it was found that by exerting sufficient pressure it could be injured in the manner described without showing any outward mark on the fish. The stomach and duodenum withstood the greatest