On the Eclipse Plumage of the Mallard (Anas Platyrhyncha Platyrhyncha)

It is a peculiarity of most species of ducks that in the early summer, shortly after the breeding season, the males moult and then assume a plumage which in coloration and form closely resembles that of the female. This is known as the “eclipse” or “summer” plumage. It lasts some weeks until a second moult, when the “normal“or “mating” plumage is again restored. During eclipse the flight feathers are replaced and the male usually deserts the female with her brood and retires either alone or with other eclipsing drakes into seclusion. The plumage changes during eclipse in various species are exceedingly intricate, and in many of the rarer species are incompletely known. Witherby (1924) publishes minute descriptions of the plumage changes of most of the British species and his descriptions are largely relied upon for the following account.

The eclipse is most noticeable in the adult males of those species in which sexual dimorphism is most marked, since the change in these cases is usually from the bright conspicuous plumage of the male into the sombre plumage of the female, but the change of plumage is not confined to these species. In the sheld-duck (Tadoma t adorna), for instance, or in the ruddy sheld-duck (Casarca ferruginea), moults take place and the male assumes female-type plumage but the differences between the sexes being slight, the change is not very noticeable. In the mallard (Anas platyrhyncha platyrhyncha (A. boschas)) and in many other species, sexual dimorphism is very marked and the eclipse is accordingly very obvious. Similarly also, although eclipse is most noticeable in the adult male, it does occurto some extent in the female, and in the juvenile birds of both sexes a partial eclipse may be shown, but here again it is not so noticeable owing to the slight change of coloration and pattern. Eclipse is therefore not to be regarded as a phenomenon affecting only the males of the sexually dimorphic breeds although most conspicuous in them.

It should also be noted that eclipse does not always imply a change of plumage in the female direction. In many cases the eclipse plumage of the adult male resembles that of the juvenile bird rather than the female, and this may apply to the whole plumage or to some regions only. When separate regions of the body are taken and when the feathers are considered individually, the changes in eclipse are most complicated. In the mallard for example the changes shown in Table I occur. The changes may therefore involve alteration of shape (tail centrals), decrease in pigmentation (under tail coverts), increase in pigment (flank), change from vermi-culated to self-coloured (flank and scapulars), increase in brown pigment (breast, flank, belly and vent), loss of iridescence (neck), increase in regional differentiation (dark streak through eye) and decrease in regional differentiation (loss of white neck ring). The very diversity of these changes suggests a profound and complicated alteration in the internal environment, involving possibly more than one hormone acting on the growing feather and probably more than one endocrine organ. Not only are plumage changes involved but also a change in behaviour. Thus Grey (1926) and Perry (1936) draw attention to the fact that the males of duck species which go into eclipse take no part in incubation or after care of the young; they leave the females and go into seclusion. Non-eclipsing ducks on the other hand share the parental solicitude. As has been pointed out above, however, it is difficult to draw a distinction between apparent absence of plumage change and a more or less complete change to the female type, except in those species where there is marked sexual dimorphism, and therefore the correlation between paternal indifference and eclipse is perhaps more correctly stated as a negative correlation between paternal solicitude and sexual dimorphism.

One other interesting feature of the eclipse may be mentioned. Perry (1936) states that the eclipse is found almost exclusively among ducks breeding in the Northern Hemisphere, and further that some of the non-eclipsing ducks of the Southern Hemisphere evolve an eclipse after transference to the North. This will be discussed later. (See p. 444)

In this paper two aspects of the eclipse problem have been subjected to experimental treatment; the first concerns the seasonal periodicity of the eclipse and the second the nature of the endocrine mechanism involved.

When these experiments started the work of Rowan (1930), Baker & Ranson (1932) and Bissonnette (1932) on seasonal periodicity suggested that light might have an important influence on the assumption of eclipse and the following experiment was devised.

On 15 December 1931 two similar experimental pens were arranged inside a brick-built poultry house. Each pen was about 10×4 ft. and was provided with a hatch which led out into a small outside yard. One pen was screened with sheets of brown paper and was lit by a 200-Watt electric light hung in the middle of the pen about 3 ft. from the ground. This light was turned on at dusk when the birds were shut up for the night and remained alight until 11 p.m. Two pairs of newly trapped wild mallards were placed in each pen. They were fed grain and some poultry laying mash and had some green food occasionally. A change in behaviour of the two sets of birds was first noticed on 3 January when the lighted drakes showed courtship behaviour which takes the form of a spasmodic dipping of the head. On 7 January the two drakes started fighting. One (No. 387) rapidly assumed dominance over the other (No. 392) vigorously driving it away from the females. On 9 January, fearing a mortal fight, the drakes had to be separated each with a female in different pens. The same lighting was continued in each pen. On 7 February drake No. 387 began to moult, followed 2 days later by No. 392. By about 20 February all lighted birds, both male and female, were moulting and the male plumage was gradually being replaced by eclipse. On 18 March the birds were photographed (see Plate I). In the photograph both lighted males show definite eclipse plumage, although not yet complete. It will be noticed that the controls were still living quite amicably together and the males are in full normal plumage. On 4 April, however, the control males started to fight and had to be separated. The lighted males were by this time in full eclipse. On 16 May they started to moult again and throughout June and July there was a gradual reversion to normal plumage. It was not until the middle of July that the control males moulted and assumed the eclipse which lasted until about the middle of September when the plumage returned to normal. These results are summarized in Table II.

It will be seen from the table that not only was the cycle of male activity (courtship, fighting and assumption of eclipse plumage) initiated much earlier in the season in the lighted birds than in the controls, but also the period between “fighting” and the first moult was much shortened. This may indicate that light not only initiates the cycle but has a continuous effect in speeding up the whole cycle. On the other hand the period of eclipse was somewhat lengthened. It must be pointed out, however, that as the artificial lighting was put on at dusk and turned off at 11 p.m. the additional lighting became gradually less as the season advanced. About the middle of August the lighting was discontinued. None of the pairs, either experimental or control, were ever seen to copulate and no eggs were laid. It was thought that close confinement of the birds in the experimental pens and continual disturbance during the day by passers-by were responsible. The following year, therefore, the experiment was repeated under different conditions; our object being to determine whether an initiating stimulus was alone sufficient for the whole cycle, whether continuous additional lighting throughout was necessary, and whether the females would lay eggs and rear them if provided with more natural conditions free from continual disturbance.

In order to accommodate the various birds which we had under observation at the time, an area of about an acre had been enclosed with 1 in. wire netting 6 ft. high. This enclosure included a stretch of brook, one bank of which was planted with trees and thorn scrub. The birds had, therefore, water, shade, privacy and natural feeding. During winter when necessary the birds were fed grain but they obtained a considerable amount of food from the brook and herbage of the enclosure. They kept remarkably fit and in good condition. Within this enclosure a small pen of netting 10 × 10 ft. was made and lit by a 200-Watt lamp hung 3 ft. from the ground in the middle of the pen. The light was regulated by means of a clock-switch situated at some distance from the pen so that the only time the birds were visited was when food and water were provided once a day. Four pairs of newly caught mallards were placed in this enclosure on 20 December 1932 and given artificial lighting from dusk until 11 p.m. as before. It was proposed to continue the lighting until the birds showed a definite response by courtship and fighting and then to liberate them in the general enclosure when it was hoped they would find conditions favourable for mating and nesting. The experiment was a failure. Although the lighting was continued until May no change was noticed and finally the experiment was abandoned and the birds liberated into the large enclosure. No explanation of the failure can be given. It was noticed, however, that the birds did not settle down in the enclosure but remained very wild and were always much frightened at the approach of the attendant with food and water. It is also possible that exposure to inclement weather and lack of shelter prevented the response. After liberation some of the birds mated and laid eggs but at about the same time as the controls which had not been lighted.

In spite of the failure of the second experiment, the results of the first are sufficiently concise to indicate that the mallard is sensitive to light and that the assumption of eclipse plumage which normally occurs in early summer is probably due in some part at least to the increasing exposure to light as the season advances since it can be greatly accelerated by additional exposure to light in early spring. The failure in the second experiment may also indicate that exposure or fright may inhibit the response. In this connexion it may be mentioned that Morgan (1919, p. 75) quotes the earlier statement of Yarrel that “a (wild mallard) male shut up by himself from early spring to the end of July undergoes no change in his plumage; but if he is allowed to associate with females till their season of incubation commences he then goes through the change and this appears to indicate the cause of the partial summer moulting”. The possibility that nutrition may also condition the effect is of course obvious. If, however, light sensitivity is a contributory cause of the seasonal assumption of eclipse, this factor must be taken into consideration when discussing the difference of behaviour between birds in the Southern and Northern Hemispheres and when transferred from one to the other, since even when climatic and nutritional conditions are similar there may be marked differences in the daily light rationing.

As one would expect the conspicuous phenomenon of eclipse has attracted the attention of the endocrinologist. Since it is most marked in the males of the sexually dimorphic species, these have been the subject of most experimentation and the fact that the phenomenon is also shown in the female and juvenile bird has been scarcely recognized. There has been much difference of opinion as to the dependence of eclipse upon the testis.

Goodale (1916) castrated Rouen ducks of both sexes. This domesticated variety of Anas platyrhyncha platyrhyncha is very similar in coloration to the mallard, shows the same marked sexual dimorphism and the male undergoes a similar if not quite so well-marked eclipse. Goodale claimed that complete bilateral removal of the testis, whether performed on young birds in juvenile plumage or on the adult in either normal or eclipse plumage, resulted in the definitive assumption of normal plumage at subsequent moultings and the birds did not show any seasonal change. Incomplete or unilateral castration did not affect the plumage changes. These results have not been confirmed by all subsequent workers. Zawadowsky (1923) found that castrated wild mallards moulted and assumed eclipse plumage in the same way as normal birds. Kuhn (1930), also using the wild mallard, found the same. Benoît (1935) and Champy (1935), however, conclude as a result of their own experience, mainly with Rouen drakes, that when castration is really complete and all traces of regenerated testicular tissue, either at the site of operation or adventitiously attached to other positions in the body cavity are eliminated, there is no seasonal assumption of eclipse.

Our own experiments were undertaken on the wild mallard.

In these experiments the birds used were wild mallards bought from commercial trappers in autumn and similar to those used in the lighting experiment. They were fully grown and in full normal plumage, but the actual age was unknown. Each bird was pinioned and provided with an identification tab placed through the wing. Castration was performed under anaesthesia produced by intramuscular injection of Numal-Roche (diethylamine-allylisopropyl-barbiturate), about 0·5 c.c. per g. body weight being sufficient to induce complete anaesthesia in about 30 min. Anaesthesia lasted for about one hour. The abdomen was opened on each side by an incision between the last two (caudal) ribs and the testis removed, as far as possible intact, by blunt dissection with forceps. Since at the time of operation the testes were regressed the operation was not difficult. Right and left sides were operated upon separately. In one case a bird died from severe haemorrhage resulting from rupture of the vena cava. In the others recovery was uneventful, except that it was noticed that when anaesthesia was very deep respiration tended to take place to some extent through the opened air-sacs, and that if the incision was sewn up while the bird was in this condition, respiration might cease altogether. One bird was lost in this way. The day following the operation, the birds were liberated into the “brook” enclosure described above, along with unoperated controls and a number of other duck species which we had under observation. The birds were disturbed as little as possible but kept under observation from day to day. At intervals of about one month in the summer and less often in the winter the birds were rounded up and caught for examination. The birds were under conditions as nearly approaching the wild as possible. This had, however, some considerable drawbacks. The birds remained very wild and were not easily caught. In spite of the wire netting enclosure and grids placed across the entrance and exit of the brook many managed to escape and some were killed by stoats or other predatory animals. Table III summarizes the relevant material, excluding those birds which were not under observation a sufficiently long time to determine whether any change had occurred or not.

It will be seen from the data that incomplete removal of the testis (No. 388) resulted in quite obvious regeneration and eclipse reappeared in two subsequent seasons. In three birds eclipse plumage was assumed during the first season but did not appear the following year. One of these birds was examined and no gonadal tissue could be found; the other two escaped before autopsy could be made. In one bird eclipse plumage was assumed in the first season but only a partial eclipse in the second; unfortunately this bird was killed accidentally and partly eaten before it was found. Most of the carcass, however, was intact and no gonadal tissue could be discovered at the site of the original gonads.

Now these results are not as conclusive as one could wish owing to paucity of data and absence of final autopsy, but the fact that eclipse appeared in the first year and not in the second year in three birds is of considerable interest. If we assume on Benoît’s and Champy’s hypothesis that the assumption of eclipse was due to minute traces of gonadal tissue left behind at operation, then it is clear that this tissue is not always capable of maintaining its seasonal activity or that in spite of its presence in early stages it gradually undergoes extinction. On the other hand it must be borne in mind that direct proof that the gonad of the duck does produce a hormone or hormones capable of acting directly upon the feather follicles and able to produce the diversity of changes which characterize the eclipse has not yet been obtained. Witschi (1935) finds in indigo buntings and African weaver finches that seasonal changes in plumage may be due to endocrine organs other than the testis, probably anterior pituitary. The possibility, therefore, remains that the eclipse is conditioned by a similar mechanism and that removal of the gonad only brings about a castration effect indirectly and more or less slowly according to the conditions of the operation and possibly according to the species or variety used in the experiment. This may explain the discrepancy between the results of various workers. Some of Goodale’s (1916) and Benoît’s (1935) own results are interpretable by this latter hypothesis. Thus one of Goodale’s castrated drakes (No. 172) went into eclipse after operation. On autopsy a small nodule of tissue was found at the site of the testis. There is, however, no evidence that this nodule was active; indeed the genital tract of the bird showed distinct signs of castration. Further, similar nodules were found in birds which did not show eclipse. Benoit (1935) records similar cases. No. F 75, castrated in April 1932, showed eclipse in the following May to July, but not in subsequent years. A laparotomy in March 1933 revealed no trace of testicular tissue. No. F 89 was castrated in April 1932. The first moult in May 1932 showed slight traces of eclipse on the head. Laparotomy in March 1933 revealed no trace of testicular tissue. No. F 90 was castrated on 22 April 1932. Partial eclipse of the head occurred at the moult in May to June. Laparotomy in March 1933 revealed no trace of regenerated tissue. Although Benoit explains these cases as due to persistence of the hormone for some time after removal of the testis, this hypothesis does not fit in with the fugitive nature of most of the hormone preparations found to affect feather coloration in other species nor can it be applied to the data presented in this paper. Kuhn (1930) also found from plucking experiments that the eclipse hormone in the intact bird could not be demonstrated by its effect on the regenerated feathers before May, i.e. just before the actual moult.

Although the actual endocrine causation of the eclipse cannot on the above data be attributed to the action of any one particular organ, there seems to be adequate reason to assume that the testis is not the sole source of the effective hormone and may be only indirectly involved. Since the eclipse is a seasonal phenomenon, sensitive to light treatment, we may legitimately expect that the anterior pituitary is somewhere concerned in the chain of causation, but as pointed out above the changes shown at eclipse are so complex that to postulate a simple direct action would be quite unwarrantable at this stage.

These experiments were undertaken at the suggestion of Dr F. H. A. Marshall, who not only gave his personal assistance and advice, but also defrayed the expenses from a grant from the Royal Society. Prof. Bissonnette kindly took the photographs for Plate I. Mr James Pike gave us much help in handling the birds. Dr Hammond and Mr Pease of the Animal Research Station kindly provided accommodation.