The Causes of Genetic Variation

In the last chapter we examined some of the evidence offered in support of the belief that adaptation in highly organised forms is a consequence of the inheritance of adaptative changes induced by the influence of external conditions. The state of knowledge of this whole subject is, as I have said, most unsatisfactory, chiefly for the reason that in none of the cases which are alleged to show a positive result have two observers been over the same ground, or as yet confirmed each other. In the wider consideration respecting the causes of variation at large we find ourselves still in the same difficulty. The study has thus far proved sadly unfruitful. In spite of the considerable efforts lately made by many observers to induce genetic variation in highly organised plants or animals, and though successes have occasionally been announced, I do not know a single case which has been established and confirmed in such a way that we could with confidence expect to witness the alleged phenomena if we were to repeat the experiment. Abundant illustrations are available in which individuals exposed to novel conditions manifest considerable changes in characters or properties, but as yet there is no certain means of determining that germ-cells of a new type shall be formed.

Of the direct effect of conditions the lower organisms, especially bacteria, offer the best examples, the alterations of virulence which can be produced in so many distinct ways being the most striking and familiar. That attenuation of virulence can be produced by high temperatures or by exposure to chemical agents, and that this diminution in virulence may remain permanent is, from our point of view, not surprising; but the fact that in many cases the full virulence can by suitable cultivation be restored is difficult to understand. Similar variations have been observed in power of pigment production and other properties.

These phenomena naturally raise the question whether any cases of apparent loss of factors in higher forms may be comparable.

The subject of variations in the lower organisms and their dependence on conditions is a highly special one, and I have no knowledge which can justify me in offering any discussion of them, but I understand that hitherto little beyond empirical recognition of the phenomena has been attempted. A useful summary of observations made by many investigators was lately published by Hans Pringsheim,[1] who enumerates the different agencies which have been observed to produce modifications, and the various ways in which these changes are manifested. One of the most comprehensive studies of the subject from the genetic point of view is that made by F. Wolf.[2] In his extensive cultivations of Bacillus prodigiosus, Staphylococcus pyogenes and Myxococcus he succeeded in producing many strains with modified properties. In most of these the modifications arose in consequence of the application of high or low temperatures or of the addition of various chemical substances to the culture-media. Some of the variations, which are for the most part in the powers of pigment-formation, persisted when the strains were returned to normal conditions, and others did not. In reference especially to the variations witnessed in the Cocci the reader should consult the critical account of variation in that group published by the Winslows,[3] where much information on the subject is to be found. The authors attempted to determine the systematic relationships of the several forms, as far as possible, by the application of statistical methods. The result is interesting as showing that the problem of species in its main features is presented by these organisms in a form identical with that which we know so well in the higher animals and plants, whatever properties be selected as the diagnostic characters. There are many types perfectly distinct and others which intergrade. Some of the types change greatly with conditions while others do not. This is exactly what we encounter whenever we study the problem of species on an extended scale among the higher forms of life.

There is now practically complete agreement among bacteriologists that the observations made first by Massini on the change in color of Bacterium coli mutabile grown in Endo's medium, associated with the acquisition of the power to ferment lactose, are perfectly reliable and free from possibilities of mistake. The work has been extended and confirmed by many workers, especially R. MÜller, who finds that this bacterium can similarly acquire and maintain the power to ferment other sugars. A careful account of the whole subject written by MÜller for the information of biologists will be found in Zts. fÜr Abstammungsl., VIII, 1912. After discussing the biological significance of the facts, he concludes with a caution to the effect that bacteria are so different from all other living things that generalizations from their behavior must not be indiscriminately applied to animals and plants.

In all work with this class of material there is obviously danger of error through foreign infection of the cultures, but there can be no doubt that though some of the "mutations" recorded may be due to this cause, the majority of the instances observed under stringent conditions are genuine.

Another and equally serious difficulty besetting work with bacteria and fungi cultivated from spores is that the appearance of variation may in reality be due to the selection of a special strain previously living masked among other strains. This possibility must be remembered especially in those instances which are claimed as exemplifying the effects of acclimatisation. Manifestly this consideration can be urged with most force when the strain which gave rise to the novelty was not raised from a single individual spore. Moreover, when once the possibility of spontaneous variation is admitted, it must be difficult to be quite confident that any given variation observed is in reality due to the novel conditions applied, and as I understand the evidence, the appearance of the mutational forms does not with any regularity follow upon the application of the changed conditions.

Researches into the variation of these lower forms will, no doubt, be continued on a comprehensive scale. So long as the instances recorded are each isolated examples it is impossible to know what value they possess. If they could be coordinated in such a way as to provide some general conception of the types of variation in properties to which bacteria, or any considerable group of them, are habitually liable, the knowledge might greatly advance the elucidation of genetic problems.

Of mutational changes directly produced with regularity in micro-organisms by treatment, the experiments with trypanosomes provide some of the clearest examples. A summary of the evidence was lately published by Dobell,[4] from which the present account is taken. The most definite fact of this kind established is that certain dyes introduced into the blood of the host have the effect of destroying the small organ known as the "kinetonucleus" in the trypanosomes. The trypanosomes thus altered continue to breed, and give rise to races destitute of kinetonuclei. This observation was originally made by Werbitzki and has been confirmed by several observers. The exact way in which this alteration is effected in the trypanosomes is not quite definitely made out, but there is good reason for supposing that the dyes have a direct and specific action upon the kinetonucleus itself, and circumstances make it improbable that in some division a daughter-organism without that body is produced, or that any selection of a pre-existing defective variety occurs.

Ehrlich has suggested with great probability that the dyes which possess this action owe it to the fact that they have the particular chemical linkage which he calls "ortho-quinoid." In outward respects, such as motility and general appearance, the modified organisms are unchanged, but their virulence is diminished. As regards the possibility of the defective strain reacquiring the kinetonucleus, Werbitzki states that in one case passage through 50 animals and treatment with dyes left the strain unaltered; but that in another case at the sixteenth passage 7 per cent. of the trypanosomes were found to have re-acquired the organ, and in subsequent passages the percentage increased, until at the twenty-seventh passage practically all had re-acquired it. Kudicke, however, in similar experiments did not succeed in causing re-acquisition by transplantation.

By the action of various drugs and anti-bodies races of trypanosomes resistant to those substances have been obtained. These breed true, at least when kept in the same species of animal in which the resistance was acquired. As to whether change of virulence is produced by passage through certain animals or not, there is as yet no general agreement.

Other changes, especially in size and some points of structure, are said to occur when certain trypanosomes proper to mammals are passed through cold-blooded vertebrates (Wendelstadt and Fellmer), and it is stated that these changes persist, but the observations have not yet been confirmed.

Experiments lately conducted by Woltereck with Daphnia are interesting as having given a definite positive result, in so far, at least, as the ova were affected by conditions before leaving the bodies of the parent individuals. The observations relate to the offspring resulting from parthenogenetic eggs. Females bearing ephippia (fertilised eggs) were isolated until the ephippia were dropped, and in this way the offspring of fertilisation were excluded. Males, of course, appeared from time to time in the cultures, but as fertilised eggs were rejected, their presence did not disturb the result. The most remarkable observations related to Daphnia longispina.

This species as found in the lower lake at Lunz had the front end of the body blunt and nearly round in profile; but on being cultivated in a warm temperature and with abundant nourishment the front end of the body became produced into an elongated "helmet," as Woltereck calls it. Experiment showed that the change was primarily due to the abundance of food, and owing to temperature in a subordinate degree.

This distinction arose as soon as the species was taken into the hothouse, but when the modified individuals were put back into the original conditions, a lower temperature and scanty food-supply, the next generation returned to their original form. After being cultivated for two years and about 40 generations in the more favourable conditions, when similarly put back into the lower temperature with scanty food the first generation born in these conditions was helmeted like the modified parents. Woltereck is of opinion that the ova were still unformed at the time the parents were put back, and the influence of the favourable conditions upon the unformed ova he speaks of as a "prae-induction." The effect never extended beyond the one generation, after which the strain returned to its original state.

The fact that the influence on the offspring was not manifested at first led Woltereck to expect that by more prolonged cultivation in the favourable conditions a further extension of this influence would be produced, but this expectation was never fulfilled, though the attempt was made again and again.

Similar experiments were made with Hyalodaphnia cucullata, which is far more sensitive to cultural influences, and in nature manifests a considerable elongation of the helmet as a seasonal modification, but the results were essentially the same as in the preceding case, no modification extending beyond the first generation born after the restoration to normal conditions.[5]

The only criticism of these extremely interesting results which suggests itself is that perhaps the original appearance of the modification was not in reality due to an accumulated effect of the conditions, but to some change in the conditions themselves which was not noticed. It is difficult to see how length of time or even the lapse of several generations could have so specific an effect on the race. It is no doubt often vaguely supposed by many that a long period of time may be necessary for the effect of climate or of other environmental conditions to be produced in an organism which does not thus respond at first. I have never been able to see any reason for this opinion nor how it is to be translated into terms of physiological fact, and I imagine that in those cases in which the lapse of time is really required for the production of an effect, the influence of the prolongation is rather on the conditions than on the organisms. The response of the organisms thus probably indicates not that the creature is at length feeling the effects because of their accumulated action on itself, but that the conditions have at length ripened.

As this sheet is passing through the press Agar has published[6] an abstract of evidence as to another comparable case in a parthenogenetic strain in the daphnid, Simocephalus vetulus. When fed on certain abnormal foods the shape of the body is changed, the edges of the carapace being rolled backwards so as to expose the appendages. The offspring of animals thus modified showed similar modification in the first, and to a very slight degree, in the second generation, though the original mothers were removed to normal conditions before their eggs were laid. In the third generation there was "a very pronounced reaction in the opposite direction." Agar suggests that the change may be due to some toxin-like substances, carried on passively by the egg into the next generation, against which the protoplasm eventually produces an anti-body.

The experiments which have been in recent years regarded by evolutionary writers as the most conclusive proof that direct environmental action may produce germinal variation are those of Professor W. L. Tower, of Chicago, on Leptinotarsa, the potato beetles. This work has attained considerable celebrity and has been generally accepted as making a definite extension of knowledge. After frequently reading Tower's papers and after having been privileged to see some of the experiments in progress (in 1907) I am still in doubt as to the weight which should be assigned to this contribution.

The work is described in two chief publications, the first of which appeared in 1906.[7] This treatise contains a vast amount of information about numerous species and varieties of these beetles which the author has observed and bred in many parts of their distribution throughout the United States, Mexico and Central America. The part of the book which has naturally excited the greatest interest is that in which Tower states that by subjecting the beetles to change in temperature and moisture, he caused them to produce offspring quite unlike themselves, which in several cases bred true.

It is much to be regretted that the author did not happen to become acquainted with Mendelian analysis at an earlier stage in the investigation. The evidence might then have been handled in a much more orderly and comprehensive way, and a watch would have been kept for several possibilities of error.

The headquarters of the genus is evidently as Tower states, in Mexico and the adjoining countries. In this region there is a great profusion of forms, some very local, some as for instance the well-known decemlineata,[8] more widely spread. The distinctions are almost all found in peculiarities of colour and pattern, and the limits of species are even more indefinable than is usual in multiform animals. Tower arranges the various types into seven groups of which the one most studied is that which he calls the lineata group. To this group belong all the forms to which reference is here made, and, as I understand, they differ among themselves entirely in size, colour and pattern. There is no suggestion of infertility in the crosses made between the several forms of the lineata group; in fact they present, like many Chrysomelidae, a good example of what most of us would now call a polymorphic species, consisting of many types, some found existing in the same locality, others being geographically isolated.

A series of experiments was devoted to the attempt to fix strains corresponding to the extremes of continuous variations. For example, those with most black pigment and those with least black taken from a population continuously varying in this respect, were separately bred; but almost always the selection led to no sensible change in the position of the mean of the population. The variations in these cases were evidently fluctuational. In some instances, however, real genetic differences were met with, and strains exhibiting them were, as usual, rapidly fixed.

Tower points out that several of the varieties (or species, as he prefers to call them) were obviously recessive to decemlineata. This is most clearly demonstrated in the case of the form called pallida, which is a pale depauperated-looking creature, with the orange of the thorax almost white and the eyes devoid of pigment.[9] This form behaved as an ordinary Mendelian recessive, breeding true whenever it appeared in the cultures, or when individuals found wild were studied in captivity. A black form which Tower names melanicum was similarly shown to be a Mendelian recessive. Wild specimens of this variety of opposite sexes were not found simultaneously in nature, and there was thus no opportunity of breeding them together, but the hereditary behaviour was seen in the F₂ generation from a melanicum found coupled with decemlineata. Experiments also occurred giving indication that a variety with the stripes anastomosing in pairs (tortuosa), was another recessive, and that a variety—called "rubri-vittata"—gave an intermediate F₁ with subsequent segregation. All these are forms of decemlineata StÅl.

Similar observations were made regarding forms recessive to multitaeniata StÅl. Of these two were thrown by multitaeniata itself, namely a form named by StÅl melanothorax, and regarded by him as a species, and one which Tower names rubicunda n. sp. The facts proving the recessive behaviour of their several forms will be found in the following places in Tower's book:

pallida, pp. 273-278.
melanicum, p. 279.
tortuosa, p. 280.
rubrivittata, pp. 280-281.
melanothorax and rubicunda, pp. 283-285.

Following this evidence of recessive nature of the six forms enumerated, Tower describes experiments showing, as he believes, that some of them may be caused to appear by applying special treatment to the parents during the "growth and fertilisation" (p. 287) of the eggs. The most striking example is that in which 4 males and 4 females of decemlineata were kept very hot (average 35° C.) and dry, and at low atmospheric pressure (19-21 inches). The eggs laid were restored to natural conditions. These gave 506 larvae, from which emerged 14 normal, 82 pallida and 2 "immaculothorax," viz., without pigment on the pronotum. The account of the rest of the experiment is somewhat involved, but I understand that the pallida, of which two only survived, behaved as normal recessives when bred to the type: also that the parents, after having laid the eggs whose history has been given, were restored to normal conditions and laid 319 eggs which gave 61 normals.

In another case normal parents laid 409 eggs in the hot and dry conditions, and on restoration to normal conditions, the same parents laid 840 eggs. Then 409 eggs gave 64 adults as follows:

The 840 eggs laid in normal conditions gave 123 normal decemlineata.

Similar experiments were made with multitaeniata and gave comparable results, the two recessives (melanothorax, rubicunda) being produced in large numbers when the parents were subjected to heat, but in this case the atmosphere was kept saturated with moisture, instead of dry, as in the previous instance. The same parents transferred to normal conditions gave normals only.

Lastly the form undecimlineata was exposed "to an extreme stimulus of high temperature, 10° C. above the average," and a dry atmosphere, with the result that from 190 eggs there emerged 11 beetles, all of the form angustovittata Jacoby, which subsequently bred true to that type (see p. 295).

In the results of these experiments, as described, there is one feature which I regard as quite unaccountable. Tower makes no comment upon it. Indeed, from the general tenour of the paper, I infer, not only that he does not perceive that he is recounting anything contrary to usual experience, but rather that he regards the result as conforming to expectations previously formed. The point in question is the genetic behaviour of the dominant normals produced under the abnormal conditions. These normals were the result of the breeding of parents declared to be at the same time giving off many recessive gametes. Some of these normals must be expected therefore to be heterozygous unless some selective fertilisation occurs. Nevertheless in every case they and their offspring are reported to have continually bred true. I allude especially to the tables given on pp. 288, 289, 292, and 293. Tower does not mention any misgiving about this result, and I think he regards himself as recounting phenomena in general harmony with the ideas of mutation expressed by De Vries. This they may be; but to anyone familiar with analytical breeding the course of these experiments must seem so surprising as to call for most careful, independent confirmation.

In 1910[10] Tower published an account of further experiments with Leptinotarsa. The work described related to two subjects. Crosses were made between three forms, undecimlineata StÅl, signaticollis StÅl and "diversa" named by Tower as a new species. The distinctions between these three depend partly on characters of the adults and partly on those of the larvae. The adults of undecimlineata and diversa have the elytra striped, but the elytra of signaticollis are unstriped. The larvae of signaticollis and of diversa are yellow, but those of undecimlineata are white.[11] Moreover, in signaticollis and diversa the black increases in the third stage of the larvae to form transverse bands which are absent in undecimlineata. The general course of the experiments shows that these differences may be approximately represented as due to the action of three factors, any of which may be independently present or absent. The stripings of the elytra and of the larvae are each due to a separate factor. As regards the distinction between the yellow and the white larvae the evidence does not prove that there is decided dominance of either colour and I infer that the heterozygotes are often intermediate.

The chief contribution which this new paper claims to make relates to differences in the results which ensue from crosses effected between these three types at different average temperatures.

We are first concerned with four experiments which I number (1), (2), (3), (4):

1. Signaticollis ? × diversa ? bred at an average temperature of 80º F. by day and 75° F. by night, gave two groups in about equal numbers. The first (49) was pure signaticollis and bred true. The second (53) was of an intermediate type, which on being bred together gave the typical Mendelian result—1 sig.: 2 intermediate: 1 div.

2. Next, as the account originally stood in the published paper, we are told that sig ? × div ? bred together at a day-temp. average 75° F. and night average 50° F. gave an intermediate only, which subsequently produced a normal 1:2:1 ratio. The two crosses were repeated eleven times with identical results.

In a further experiment (3) signaticollis ? × diversa ? were bred under the same conditions as those used in expt. (1). They again gave sig. and intermediates as before in fairly equal numbers. The sig. as before bred true, and the intermediate gave 1:2:1, all exactly as in expt. (1).

In expt. (4) the same parents used in (3) were again mated under conditions of expt. (2) at the lower temperature, and this time gave signaticollis exclusively, which bred true for four generations. This experiment was repeated seven times with uniform results.

Diagrams are given representing all these histories in graphic fashion.

From these observations, Tower concludes that the determination of dominance, and the ensuing type of behaviour, is clearly a function of the conditions incident upon the combining germ plasms.

It will be observed that expts. (1) and (3) gave identical results but (2) and (4), though much the same conditions were applied, are at variance, for (2) gave all intermediates, while (4) gave all signaticollis. In Amer. Nat., XLIV, 1910, p. 747, Professor T. D. A. Cockerell commented on this paper of Tower's and pointed out that there must be an error somewhere, for when he discusses these experiments Tower speaks of (2) and (4) as confirming each other. To this Tower replied[12] that there had been a mistake. He states that in preparing the paper "certain minor experiments were taken from a larger series and combined to illustrate a general point in the behaviour of alternative characters in inheritance," and that expt. (2) was introduced inadvertently in place of another which he desires to substitute. In this, which I number (5), signaticollis ? × diversa ? from exactly the same stocks as those used in (1), were mated at the lower temperatures specified for (2), day average 75° F., night average 50° F. These gave all of the signaticollis type with a narrow range of variability, which bred true, in some cases to F₆. Tower says he has repeated this experiment six times with identical results.

Nevertheless he proceeds to say that the description of expt. (2), which was repeated eleven times with identical results, was correct "as far as given." That experiment was "from a second series of cultures parallel to the one given, but in which there are other factors involved, which in H. 410 [my (2)] are productive of a typical Mendelian behaviour." He adds he does "not care at this time to make any statement of what these factors are, nor of their relations to the behaviours given in the H. 409, H. 411, H. 409/11 series [my (1), (5) and (3)—(4)] which are the simplest and most easily presented series obtained in the crossing of signaticollis and diversa."

Professor Cockerell's intervention has thus elicited the fact that we have as yet only a small selected part of the evidence before us, even as concerning the effect of temperature on the cross between signaticollis ? × diversa ?. We learn that at the lower temperatures the result was eleven times the expected one, and six times an unexpected one; further, that we owe it to the author's inadvertence that we have come to hear of the expected result at all, and that though he knows the factors which determine the discrepancy, he declines for the present to name them. In these circumstances we can scarcely venture as yet to estimate the significance of these records.

The paper goes on to recount somewhat comparable, but more complex instances in which the descent of the colour of adults and of larvae was affected by temperature in crosses between undecimlineata and signaticollis. As they stand the results are very striking and unexpected, but I think, in view of what has been admitted respecting the former part of the paper, full discussion may be postponed till confirmation is forthcoming.

One feature, however, calls for remark. This second paper is written apparently without any reference to the discoveries related by Tower in his previous book, to which no allusion is made. This is most noticeable in the case of an experiment in which (p. 296, H. 700A) undecimlineata ? (the dominant) was mated to signaticollis ? with the result that all the offspring were undecimlineata and bred true to that type (Parthenogenesis was tested for, but never found to occur). This experiment was made at a temperature averaging 95° F. ± 3.5° by day and 89° F. ± 4.8° by night, and in a humidity given as 84 per cent. by day and 100 per cent. by night; but in the previous book (p. 294) we are told that pure undecimlineata bred together "under an extreme stimulus of high temperature, 10° C. above the average" and a relative humidity of 40 per cent. gave 11 beetles only, all angustovittata. But reference to the Plate 16, Fig. 2, shows that angustovittata must be exceedingly like signaticollis, having, like it, the elytral stripes obsolete, and if there is any marked difference at all, it can only be in the larvae. It seems strange that if undecimlineata really gives off ova of this recessive type at high temperatures, the fact should not be alluded to in connection with expt. H. 700A, where, as the father was signaticollis, having the same recessive character, their appearance might have been expected not to pass unobserved. The temperature in the older experiment is, of course, not given with the great accuracy used in the second, and it may have been higher still. The humidity also was widely different. Still, in discussing the phenomena we should expect some reference to the very remarkable and closely cognate discovery which Tower himself had previously reported in regard to the same species.[13]

The hesitation which I had come to feel respecting these two publications of Tower's has been, I confess, increased by the appearance of a destructive criticism by Gortner[14] who has examined the parts of Chapter III of Tower's book, in which he discusses at some length the chemistry of the pigments in Leptinotarsa and other animals. As Gortner has shown, this discussion, though offered with every show of confidence, exhibits such elementary ignorance, both of the special subject and of chemistry in general, that it cannot be taken into serious consideration.

Some observations made by Dr. W. T. Macdougal[15] have also been interpreted as showing the actual causation of genetic variation by chemical treatment. Of these perhaps the least open to objection were the experiments with Raimannia odorata, a Patagonian plant closely allied to Oenothera. The ovaries were injected with various substances and from some of the seeds which subsequently formed in them a remarkable new variety was raised. This varying or mutational form was strikingly different from the parental type, with which it was not connected by any intergradational forms, and it bred true. It made no rosette, growing to a much smaller size than the parent, and was totally glabrous instead of being very hairy as the parental type is. I was shown specimens of these plants by the kindness of Dr. Britton in the Bronx Park Botanic Garden in 1907 and can testify to their very remarkable peculiarities. They had a somewhat weakly look, and might at first sight be thought to be a pathological product, but they had bred true for several generations. From the evidence, however, I am by no means satisfied that their original appearance was a consequence of the treatment applied. This treatment was of a most miscellaneous description. Two of the mutants came from an ovary which had been treated with a ten per cent. sugar solution. Ten came from one into which a 0.1 per cent. solution of calcium nitrate had been injected. One was from a capsule which "had been exposed to the action of a radium pencil." Macdougal speaks of these results as decisive, but clearly before such evidence can be admitted even for consideration it must be shown by control experiments that the individual plants which threw the mutant were themselves breeding true in ordinary circumstances. Nothing is more likely than that the mutant was an ordinary recessive. I may add that Mr. R. H. Compton made a number of experiments with Raimannia odorata, raised from seeds kindly given me by Dr. Britton, injecting the ovaries with a variety of substances, including those named by Macdougal; but though a numerous progeny was raised from the ovaries treated, all were normal. Macdougal relates also that some mutational forms came from ovaries of Oenothera Lamarckiana exposed to radium pencils, and also from Oenothera biennis injected with zinc sulphate a peculiar mutant was raised, but taking into account the frequency of these occurrences in those species, he very properly regarded this evidence as of doubtful application. In a later paper,[16] however, he has returned to the subject and affirms his conviction that the appearance of a mutant among seedlings raised from an ovary of Oenothera biennis treated with zinc sulphate was really a consequence of the injection, saying that the variation previously observed in the species was afterwards shown to be due to fungoid disease. The circumstances to which he mainly points in support of his view is that the mutation bred true, but this is only evidence of its genetic distinctness, which may, of course, be admitted by those who remain unconvinced as to the original cause of its appearance. He adds that he is making similar experiments with some twenty genera; but what is more urgently needed is repeated confirmation of the original observation. When it has been shown that this mutation can be produced with any regularity from a plant which does not otherwise produce it on normal self-fertilisation, the enquiry may be profitably extended to other plants.

A curious and novel experiment, which however, led ultimately to a negative result, was made by F. Payne. Many discussions have been held respecting the blindness of cave animals. The phenomenon is one of the well-known difficulties, and most of us would admit that the theory of evolution by the natural selection of small differences does not offer a really satisfying account of it. Those who believe in the causation of such modifications by environmental influences and in their hereditary transmission make, of course, the simple suggestion that the darkness is the cause of the loss of sight, and that disuse has led to the reduction of the visual organs. Payne bred Drosophila ampelophila, the pomace-fly (which is easy to keep in confinement, fed on fermenting bananas), for sixty-nine generations in darkness. At the end of that period there was no perceptible change in the structure of the eyes, or in any other respect. The number of generations may possibly be regarded as insufficient to prove anything, but comparing them, as he does, with the generations of mankind, we see that they correspond with a period of about two thousand years, an interval far longer than those which many writers in particular cases have deemed sufficient.

In his first paper Payne states that, though no structural difference could be perceived, the flies which had been bred in the dark reacted less readily to light than those which had been reared under normal conditions, and he inclined to think that the treatment had thus produced a definite effect. After more careful tests, however, he withdrew this opinion. It proved that both individual flies and individual groups of flies, both of those bred in the light and of those bred in the dark, differed greatly in their reactions, which were measured by counting the time that it took for a fly to travel to the light end of a covered tube, various sources of error being eliminated. He found further that these differences of behaviour were not inherited in any simple way, but he is disposed to attribute them to accidental differences in the nature of the food, an account which seems probable enough.[17]

In several recent publications Blaringhem[18] has described the origin of many abnormal forms of plants, especially of maize, which he attributes to various mutilations practised upon the parents. Respecting these the same difficulty which has been expressed in other cases reappears, that before drawing any conclusion as to the value of such evidence we require to know that the plants treated belong to a really pure line, which if left to nature in the ordinary circumstances of its life in that locality would have had normal offspring. Abnormalities abound in the experience of everyone who examines pans of seedlings of almost any species of plant, and in maize they are well known to be exceptionally common. Some of those which we meet with when we attempt to ripen maize in this country are very similar to those which Blaringhem describes, consisting in irregularities in the distribution of the sexes, in the shapes of the panicles, etc. Many of these are doubtless imperfections of development, due to the dullness of our climate, but others are presumably genetic and would recur in the offspring however treated. If some one working in a climate where maize could be raised in perfection would repeat these experiments, and show that a strain which was thoroughly reliable and normal in its genetic behaviour did, after mutilation, throw the miscellaneous types observed by Blaringhem, that would be evidence at least that the development of the seed could be so influenced by injury to the parental tissues that its properties were changed. Such evidence could be used for what it is worth; but pending an inquiry of this kind I am disposed to regard these observations of variation following on parental injury as suggestive rather than convincing.

Some evidence of a remarkably interesting kind has been collected by J. H. Powers[19] respecting the structure and habits of Amblystoma tigrinum, which led him to the conclusion that striking differences in the form, anatomy, and developmental processes could be effected directly by change in the conditions of life. It is well known that a profusion of forms, distinct in various degrees, is grouped round Amblystoma tigrinum. Some of these are believed to be geographically isolated, others occur together in the same waters, and, as usual, authorities have differed greatly as to the number of names to be given. These forms were studied in detail by Cope who described them in the Batrachia of North America. The view which he inclined to take was that the individual variations of Amblystoma tigrinum resulted from variations in the time and completeness of the metamorphosis, and these were regarded as due to external causes, such as differences in season, temperature, and geographical conditions. Powers, however, states that collecting within a radius of six or eight miles he found almost if not quite the whole "gamut of recorded variation in this species." Some, however, as he states, occurred rarely except under experimental conditions, but considerable differences in temperature were not found necessary in producing them. Every year, he says, he has been able to add to the number of peculiar types found in the same small area in nature, until the amount of natural variation at least equals that seen by Cope in the collections of the National Museum and those of the Philadelphia Academy.

Powers states that his observations by no means confirm Cope's view that these differences are in the main referable to variation in the completeness of metamorphosis, and on the contrary, he regards metamorphosis as on the whole a levelling process, tending to obliterate diversity. The enormous differences in size and proportions which he describes can only be appreciated by reference to his figures. They affect almost all features of bodily organisation. These striking differences he looks upon as brought about by differences in nutrition, "diversities in habitual locomotion," and diversity in the age at which metamorphosis occurs, and to sexual difference. Apart from sexual difference he regards the chief distinctions, in brief, as "acquired variations of the larva."

As an example he gives the great elongation of some of the forms as "due first to slow growth, second to the free-swimming habit, third to the prolongation of larval life, and finally to the assumption of sexual maturity as males," either in the branchiate or non-branchiate condition. He describes the rapid growth of some and the slow growth of others. A larva of intermediate type may grow about a centimeter a month, but a rapidly growing specimen may grow more than four times as much. The slower rate of growth may, he says, be induced by winter feeding, and other treatment.[20]

When, however, he goes on to describe the influences which he regards as exerted by the habit of freely swimming, I am led to wonder whether after all in most of these illustrations, the primary distinctions are not in reality genetic. "Specimens raised in the same aquarium or in similar aquaria, side by side with all conditions as uniform as it is possible to make them, seldom fail to furnish striking examples of broad-headed, short-bodied, and short-tailed types which are habitually found at the bottom, while others, slender and elongated, are free swimmers, and maintain themselves in almost as continual suspension and motion as does a gold fish." Later, again, he writes, "Yet despite the uniformity of these favourable conditions, the larvae soon began to split up into two noticeably distinct groups, the one of unusually compact proportions, the other of uniform intermediate build, such is most commonly met with." It is to my mind scarcely possible to resist the inference that, though there may be definite responses to certain conditions, yet the chief distinctions are genetic, and that it is these distinctions which confer the power to respond. The parts respectively played by cause and effect are always difficult to assign; but when it is stated that "a weak-limbed, long-bodied and long-tailed animal becomes well nigh perforce an undulatory swimmer, while the strong-limbed, short-tailed, heavy-bodied specimen, when these characteristics are rapidly forced upon it, is, under certain circumstances, just as forcibly induced to become a crawler," we feel how erroneous any estimates of causation are likely to be.

One of the most remarkable and interesting sections of Powers' paper is that in which he describes the differences in bodily structure and habits which he attributes to cannibalism, and the whole account of the phenomena should be read in the original. It appears that there are two extremely distinct types of larvae, those with narrow heads and slender bodies which live for the most part on small Crustacea such as Daphnias, and those with huge mouths and very wide heads, which disregard such small animals altogether and live on amphibian larvae, whether of their own or other species. As the illustrations show, the differences between these two types are very great, and the differences in instinct and behaviour are no less. The cannibals take no heed of the pelagic crustacea, lying sluggishly at the bottom, rousing themselves immediately to a violent attack on the larger living things which approach them. Nothing but the most incontrovertible evidence based on abundant control experiments should convince us that such differences are not primarily genetic, and in the present state of knowledge I incline to think that the families really consist of individuals which are ready to assume the cannibal habit if opportunity offers, and others which are congenitally incapable of it. It may readily be that if all chance of cannibal diet be excluded, the full development of the wide head and mouth, or the other peculiarities, would never become pronounced, but I doubt whether such change could be induced in any individual taken at random.