Materials for the Study of Variation Treated with Especial Regard to Discontinuity in the Origin of Species

1 Partial deficiency of scales, occurring evenly over all the four wings, is not very rare in S. carpini. I have myself reared two such specimens. and irregularly shaped patch of pale colour. If this specimen, which was in" the collection "f the late Mr F. JJond, is still in existence it is greatly to be wished that a proper description of it should be published.] 460. Saturnia carpini .' : \vings yellowish-grey throughout, with the usual markings, save that on the fore wings there is no ocellus, and on the hind wings is only a small black eye, without a border, having a yellowish-givy central spot. OcHSENHEIMER, F., Schmet. von Europa, 1810, iv. p. lit I.

From this evidence it is clear that the range of Variation of ocellar markings in Lepidoptera is very great. It is especially to be noticed that this variability affects no one family, or the species of one geographical region, or one kind of ocellus exclusively, though doubtless it is more marked in some than in others ; but it seems rather to be a property belonging to ocelli in general. From the fact that they can bodily come and go, it seems clear that, as was suggested above, each ocellus is as regards its origin one structure made up of parts in correlation with each other.

The great variability of ocellar markings is probably not peculiar to Lepidoptera, but I have no evidence sufficient to produce regarding the variability of ocellar markings in other forms. I may however instance the case of the Raiidae, many of which have been found marked with a large ocellar mark on the dorsal surface of each pectoral fin. At different times such a mark has been thought to characterize a certain species, but I believe it is now generally admitted that it may appear as a variation in several species. The best figure of this ocellar mark is that given by DONOVAN (Brit. Fishes, 1SOS, v. PI. CHI.) in a Ray described under

461. the Linnean name Raia mindetas. On each "wing" was a large spot, having a dark purple centre, surrounded by a zone of silvery green enclosed by a broad dark boundary composed of five equidistant, contiguous spots of blackish purple. Donovan suspected that the fish might be a variety of the Homelyn (R. maculata), and it has been generally believed by other authors to have been so. Donovan states that a similar eye-spot was seen by him in various degrees of definition in several young Skates.

462. R. clavata, the Thornback, also sometimes has a large white spot surrounded with black on the " wings." DAY, Brit. Fishes, II. p. 344. Raia circularis, the Cuckoo Ray, has normally on each "wing" a large black blotch banded with yellow and surrounded by yellow spots. This structure may be absent as a variation. DAY, Brit. Fishes, II. p. 349. Reference was made (Introduction, Section V.) to that relation subsisting between the several members of a linear series of segments or other repeated parts, by virtue of which they m;i\ resemble each other in respect of colour or pattern of colours. From the fact that the several members do in such cases often bear the same colours or patterns it is clear that they must at some time or other have undergone similar Variation. In order to measure the possible rapidity of the process of evolution by which such parts may have reached their present condition it is important to ascertain the extent to which their several variations may be simultaneous.

Variations in colour are of course Substantive variations and a full consideration of their nature cannot be taken here. For the present we are only concerned with the consequences of the fact that the parts are repeated in series. As was pointed out in the Introduction the problem of the resemblance between the colours of such segments is only a special case of the same problem of Symmetry which is again presented in bilateral or other Repetition.

Simultaneous colour-variation taking place abruptly in a large number of organs, such as hairs, feathers, &c. is a very common occurrence, and the part that repetition of structures plays in producing the total effect is apt to be overlooked. In comparing two varieties of some whole-coloured animal, a bay horse with a chestnut for example, it must be remembered that the difference is really made up of a simultaneous variation in the pigment of each particular hair. Similarly if a caterpillar normally green appears in a uniformly brown variety we may conceive the total change as brought about by variation occurring simultaneously in the skin of the several segments, or in some smaller units. But whatever unit be taken, whether segment, or hairs, or cells, that all or any particular groups of such units should vary together and in the same direction is not a matter of necessity. That such simultaneity is not universal and that segments may vary independently of each other is a matter of common observation, and indeed is sufficiently proved by the occurrence of differentiation between segments. Nevertheless the evidence goes to shew that between parts repeated in series there may be a relationship of the kind spoken of, though its causes, nature and limitations are unknown. In the case of actual segmentation this relationship may appealeither in the simultaneous variation of the colour-patterns of the segments, or of some one colour or patch borne by each, or by the appearance of some unusual mark or patch on several of them at once.

In some cases it happens that certain of the segments may vary together, the rest remaining unchanged, and, as seen in Chiton I/HI nnoreus, (q.v.}, the segments thus undergoing the same variation are not always even adjacent to each other. The win iltquestion is a very large one and it is not possible here to do more than refer briefly to a few cases illustrating some of its different aspects. Fuller treatment will be attempted in connexion with the evidence of Substantive Variation.

4G3. -^s '-Mini] iles of a form whose segments in their colourvariations manifest a very close agreement with each other, the Hirudinea may be taken. Figures of numerous varieties of medicinal Leeches are given by EBHARD. Xnm-i-II,' nnnioijr. des Sangsues, 1857, and other cases are represented by MoQUlN-TANDON, Monogr. de la famiUe des llirudinees, \*'1~ (see especially PI. v. tig. 1). As these figures testify, there is a widediversity both in the ground-colour and in the size, colour and manner of distribution of the lines and spots with which it is decorated, but M> far as may be judged from the figures and descriptions the same decorations are repeated on the various segments. It cannot be doubted that a close scrutiny of the specimens would shew points of difference even between adjacent segments but substantially the patterns are the same for the segments of an individual. The patterns of the varieties may thus, like patterns of ribbon, be each represented by a drawing of a short piece of the body in the way adopted by the writers named.

As regards the larv;e of Lepidoptera a good deal of information 1 tearing on this subject exists, and some of these results, especially those relating to Sphingidse, are of interest1. *464. ^n ^ne kirvjv °f "iany species of Sphingidse there is a more or less regular dimorphism in colour. Notable examples of this are AcJterontia atropos, Ch(erocan//»t elpenor, C. porcellns and M<t<-r<>'jl<>**<i xti'llittitrnin, in each of which the larva is known both in a light green and in a dark form". The dark form is the commonest in C. porcellus but in A. <tfr<>/it>x it is much rarer than the green form. Judging from the figures, the ground-colour of the segments generally varies as a

'• That this dimorphism is ' phytophagic' is not very likely, but the possibility should be remembered. It seems to be established that in many of the species the colourvarieties are definite and largely discontinuous. Of M. xtellutarum WKISMAXN (p. 250) bred 140 from one batch of eggs, and of these 49 were of the green form and 63 of the brown form, only 2s bring transitional. The discontinuous character of the variation was illustrated by one most remarkable specimen. In it the body was /Hirtii-oltiun'd. being ]i!irtly of the green and partly of the brown form. The bead, protborax, all the abdominal segments behind the 2nd, and the right side of the remainder were brown, but the left, side of the meso- and meta-thorax, of the 1st abdominal, and part of the left side of the 2nd abdominal were green [according to the figure 9 on PI. in., with which the description in tbe text, p. '_' I'.), differs slightly). In .1. iitrviHix I know no account of any intermediate form. In most of the species the dimorphic or polymorphic character appears in tbe later periods of larval life and especially after tbe last moult; but in ('. /lori-i-llux, according to both WKISMANN (p. 188) and BUCKLER (p. 117) though the lame are of both kinds in the penultimate state all or nearly all after tbe last moult turn to the dark form.

whole, shewing only slight differences in tint in different parts of the body. To this there are certain exceptions, of which A. atropos is especially remarkable. In the brown variety of this species the abdominal segments have a dark ground-colour composed of shades of brown, while the three thoracic segments in it are white "like linen" (see WILSON, PI. vi.; BUCKLER, PI. xxi.; POULTON, 1886, p. 149; HAMMOND, Zool., 6282; BALDING, Ent. Mo. Mag. xxn. p. 279; GIRAKD, Bull. Soc. ent. Fr., 1865, S. 4, v. p. xlix. <fec.).

Tn M. stellatarum though the ground-colour of the head and of all the segments varies greatly it appears that the head and prothorax vary in colour simultaneously with each other and are of one colour, while the other two thoracic segments and the abdominal segments also vary together but usually differ from the head and pro-thorax (see WEISMANN, PL in.). In illustration of the degree to which simultaneity of Variation is possible over considerable areas of the body the varieties in markings are perhaps more important than those in ground-colour. Of such changes simultaneously occurring in several segments there ai'e many examples.

4g5 In all the varieties of ground-colour in M. stellatarum the pattern of the markings remains the same though of differing intensities (WEISMANN, p. 248), but in the brown variety of A. atropos the pattern is quite peculiar and cannot even be recognized as a representation of the markings seen in the green form. Even the oblique stripes are absent (PouLTON, 1886, p. 149; see also authors quoted above). But as in the ground-colour so in the markings, the abdominal segments have one new pattern while the thoracic segments have another.

466 The figures of larvte of Deiphila euphorbia} given by BUCKLER and by WEISMANN are especially interesting in this connexion, shewing that in the complex variations of this polymorphic form the particular pattern of the individual is carried out with little difference in each segment behind the prothorax. Some of these changes are extensive, but to be at all appreciated the figures must be referred to. In one case all the triangles at the posterior part of each segment were red instead of green as usual, and this change was found in many specimens from one locality (see WEISMANN, p. 206, PI. v.). This identical variation was known to and figured by HUBNER (WEISMANN). In one specimen from the same place as the last the second row of marks which should occur just below the sub-dorsal mark of each segment was absent throughout the whole line, and the ring-spots of the upper or sub-dorsal row had, as a variation, a red centre or nucleus, well marked in the posterior spots but fading away anteriorly. The occurrence of these considerable changes is still more noteworthy if, as WEISMANN states, the members of each batch are much alike. He remarks also that the variability is great in some localities but little in others.

467. The larva of Deilephila hippophaes has a sub-dorsal row of red markings upon a variable number of segments from the 7th abdominal to the 3rd or even 2nd abdominal, increasing in size and distinctness from behind forwards. The size of these markings differs greatly in different specimens, varying from a mere dot to a distinct red spot with a black ring. As the figures shew, there is a considerable correspondent Ix-twccn the segments in the extent to which the spots are developed, though in each case they fade away in the anterior segments (see \\'I.I~M.\N\'S tigs. 59 and 60).

Another interesting example of considerable uniformity in the colour-variation <>i a series of segments is to be seen in Saturnia earpini. In this species besides change in the tint of the green _'i »und-colour [two chief tints being found, one dark and one light] there is immense difference in the amount of black pigment deposited, mo.-t marked in the last two stages of the larvse. Good figures and descriptions of these are given by WEISMANN (PI. vin.). Though no two .segments are alike and though there are differences perceptible even bet ween the two sides of most segments, yet the general scheme of colour of eaeh individual is carried out with fair constancy over the several segments. As I have myself seen, the lightest and darkest may both be reared from one batch of eggs and in the same breedingcage or slee\ e.

469. The colour of the tubercles of S. earpini also varies greatly. They may be light yellow, dark yellow, pink, violet, or white, but the yellow and pink forms are the commonest. As I have myself observed, there is generally a close agreement between the different tubercles of each larva in point of colour. In a few specimens I have seen the tubercles of the anterior and posterior segments pinkish, while the remainder wen? yellow, but this diversity is exceptional. The importance of this case is increased by the fact that POULTOX (1887, p. 311) has found that the offsjiriiii: of a pair whose tubercles had been pink shewed a high proportion of larva' with pink tubercles. The two parents were from a lot of 80 larvae found together, of which only 3 had pink tubercles : but of their 88 offspring 64 had pink tubercles.

470. The case of the occurrence of red spots on the larvag of Smerinthus ocet/dtiix and S. pojmli1 may be quoted as an instance of great irregularity in the degree to which the segments agree in their colourvariations. This well-known case is also of great interest as an example of a parallel variation occurring in different species. The larva1 of both species are most commonly without any red spots, but not rarely a number of red spots are present. In extreme cases each of the spiracles is surrounded with red, and there is in addition a row of red spots in the sub-dorsal region of all segments from the 1st thoracic to 7th abdominal, and also a red spot on each clasper. The number of spots, number of rows, the size and tint and distinctness of the spots is exceedingly variable. In point of time the spots of the 3rd abdominal segment appear first and those of the 2nd thoracic next (POULTOX, 1887, p. 28f>, Arc.). Though in much spotted specimens the spots may remain till the larva is full-fed, in some cases a few spots appealat an early stage and are afterwards lost. Among the individuals of the same brood there may be great diversity, some having spots and others being without them (PouLTON, 1887, p. 287). In several cases a spot present on one side of a segment has been found absent on the other side. As Poulton observes, it is especially

1 I have not referred to the case of S. tilia, as it is possibly of a different nature. remarkable that though there are no spiracular openings on the rneso- and meta-thoracic segments, yet in cases of extremely spotted larvae there are red spots at the level of and continuing the spiracular series of spots upon these segments also («S'. ocellatus, BUCKLER, PI. xx. tig. 1 a; POULTON, 1887, PI. x. fig. 1. S. populi, POULTON, 1887, p. 286). As an indication of an element of detiniteness in this variation may be mentioned the fact that in fully spotted larvee of S. populi the sub-dorsal spot on the 7th abdominal seems to be always the smallest in that row (POULTON, 1887, p. 285; WILSON, PI. v. fig. la; FLEMYNG, Ent., 1880, p. 243, etc.).

In our present consideration the fact that these very large variations sometimes occur simultaneously over a large range of segments and are sometimes restricted to particular segments is of considerable importance. We may note that WEISMANN (p. 360) is prepared to believe that these spots represent a new variation arising similarly and independently in the different species of Smerinthus. As however is usual in cases of considerable Variation an attempt has been made to lessen the value of these indications of the magnitude of Variation by suggesting that they may be of the nature of " reversion " (PouLTON, 1884, p. 28). Apart however from a general reluctance to recognize the possibility of the occurrence of large variations there seem to be no special grounds for the suggestion here. It is nevertheless true that in the case of the Smerinthus larvse a complete disproof of the hypothesis of " reversion " is wanting. This is only to be obtained in cases (like that of D. euphorbia?}, in which a great number of complex and mutually exclusive variations exist side by side. In the absence of such complete refutation the hypothesis of reversion may still find favour.

*471. Chitonidae. The following facts observed in certain Chitons are given in illustration of the existence of a similar possibility of simultaneous Variation between parts which are repeated in series but whose repetition is not of the kind commonly included in the term Metameric. Unfortunately the material at hand is very limited and I do not know what might be the result of further examination, but the facts seen suggest that the subject is worth investigating.

The dorsal plates of Chitons are eight in number. Though the colours and markings in different species are complex and various yet in many species all the plates are alike or nearly so. The question then arises do all the plates change colour together, or do they change one by one, or otherwise ? From the few observations made it seems that in this respect the species differ, but variation uniformly occurring in all the plates seems to be rare. This may perhaps be due to the constitution of such specimens as separate species, but I saw little likelihood of this. On the other hand in several cases the same variation was present in more than one segment, and in particular there was strong evidence that in some species the segments 2, 4 and 7 shew a noticeable

agreement with each other in colour-variation. The specimens are all in the Mat-Andrew Collection in the Cambridge University Mii-einn, and I have as usual simply followed the labelling of the specimens. C. ui-l'itxttini, An-tialia. 10 specimens, of which the plates in 6 are nearly uniform. In one there is a white hand in the centre of each plate; in 2 the plates are irregularly coloured; in one the plates 1 and 0 agree in being broadly marked with whitr.

( '/til', n lii'iimilii, IVru. 4 specimens. 3 are uniformly dark brown; but in the other specimen there is a .-//•«/<;/ u-ltitt' mark on the centre of plates 2 — 7, and a faint one on plates 1 and 8. C. elegant, Chili. 2 specimens. In one, complicated markings are repeated on each plate nearly uniformly; in the other specimen a much simpler pattern recurs on , iir/i .-I'liiin nt. On the other hand, C. pellis-serpentis, New Zealand, 8 specimens: great diversity of markings and no uniformity among plates in 4 specimens, but in one specimen plate* 2 — 5 were black and the rest light-coloured. Similar want of uniformity among the plates in 2 specimens of C. incanus, New Zealand.

The evidence of agreement between segments 2, 4 and 7 in the following cases is very .strikm-. C. I Tniiifini iimriiinri'iia, "Hebrides, <$rc." 18 specimens, all of a light brown colour marked with dark red. In 1 specimens the plates are uniformly marked or nearly so. In 6 specimens plates 2, 4 and 7 are much darker than the others, being for the most part of a uniform dark red. In 5 specimens plates 2, 4, 7 and 8 are darker than the rest. In 1 specimen plates 2, 4, 5 and 7 are darker than the rest.

In 2 specimens the central parts of most of the plates have dark markings, but no segment is specially distinguished. Of 18 specimens therefore 12 have plates 2, 4 and 7 darker than the rest. Among 3 specimens of the same species from Gr. Manan (N. America) 2 are nearly uniform throughout, but iu one plates 2, 4, 7 and 8 are much darker than the rest. C. (Toniciit) liwatus, 2 specimens. In one the markings on all the plates are nearly similar, and the white wavy stivaks characterizing the species are almost similarly distributed on the sides of all the plates. In the other specimen these lines are absent on the plates 2, 4 and 7, which are much darker than the rest; but the lines, though less extensive than in the first specimen, are present on plati-> 1, :?, 5, 6 and 8.

The preceding evidence may suffice to indicate the nature of this important question of the degree to which the colour-variations of parts repeated in Linear Series may be similar and simultaneous, a question which, as must be evident, is of the highest consequence in estimating the magnitude of the steps by which Evolution may proceed. To the consideration of this matter it will be necessary to return when the evidence of Substantive Variation is considered.

Meanwhile it, will not be forgotten that though we have only spoken of this question in reference to colour and to Linear Series, t he same question arises also with regard to other variations and in reference to all parts which are in any way repeated and resemble each other, whether such repetition is strictly serial or not. In a survey of any group of animals cases will be seen in which organs in one region are repetitions of organs in another region though

not necessarily in serial homology with them in any sense in which the term is commonly used. Many such cases were spoken of by Darwin in the chapter on " Correlated Variability1" and are now famous. The simultaneous colour-variations of the mane and tail of horses2, the correspondence between the large quills of the wings and those of the tail of pigeons3 and other birds are among the most familiar of such cases. When with such facts in mind we turn to some species which differs from an ally in the presence of some characteristic development or condition common to a number of its parts, in making any estimate of the steps by which it may have been evolved it must be remembered that it is at least possible that the common feature characterizing these several parts may have been assumed by all simultaneously. To take a single instance of this kind, the species of the genus Hippocampus, the Sea-horses, have the shields produced into more or less prominent tubercles or spines. The back of the head is also drawn out into a prominent knob. In an allied genus from Australia, Phyllopteryx, many of these spines are provided with ragged looking tags of coloured skin, like the seaweed which the fishes frequent4, giving the animal a most fantastic appearance and no doubt contributing greatly to its concealment [probably from its prey]. If in this case it were necessary to suppose that the variations by which this form has departed from the ordinary Hippocampi had occurred separately, and that each spine had separately developed its tag of skin, the number of variations and selections to be postulated would be enormous ; but probably no such supposition is needed. We are, <is I think, entitled to expect that if we had before us the line of ancestors of Phyllopteryx, we should see that many and perhaps all of the spines which are thus modified in different parts of the body had simultaneously broken out, as we may say, into tags of skin, just as the feathers of the Moor-hen (Gallinula chloropus)5 may collectively take on the " hairy " form, or as, to take the case