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

*63u. Sarsia mirabilis ' : normally four radial canals, &c. (Fig. I'll. 1 and 111). Out of many hundreds of N.American specimens In',, were t-Mind with six radial canals, six ocelli, and six tentacles, FIG. 127. Surain miniliili.*. I and III, the normal form, with four radii, from mill from uliove. II and IV, au abnormal form with .six radii, from below and from ubuvu respectively. (From AGASSIZ.) coast of Scotland one was found having six radial canals, six

xn. p. 527. *632. Sarsia sp. A single specimen having Jive complete segments: more striking case of the Discontinuity and perfection of Meristic Is it besides a mere coincidence, that the specimens presenting this variation, so rare in the free-swimming Hydromedusse, should 633. Clavatella (Eleutheria) prolifera. This form has a medusa which creeps about on short suctorial processes borne by the tentacles. The number of these tentacles varies from 5 to 8. In the specimens examined by KROHN' the number was 6. Most of CLAPAREDE'S* specimens had 8. FiLiPPi3 found that the majority had 6 arms, but 15 per cent, had 7. Those examined by HiNCKS4 never had more than 6. Filippi considered that the difference in number was evidence that his specimens were of a species different from Claparecle's. I examined many of this form at Concarneau and found six the commonest number in the free medusas, but those still undetached frequently had 5, possibly therefore the number increases with development. [See also Cladonema radiatum, &c. HINCKS, I.e., p. 65, &c.]

Claparede states that the 6-armed specimens had 6 radial canals, but the 8-armed usually had four though occasionally six, but never eight canals. In this case note not only the frequent occurrence of Meristic Variation, but also the suggestion that particular numbers of tentacles are proper to particular localities. *634. Normally there is a single eye at the base of each arm. CLAPAREDE figures (I.e. p. 6, PL i. fig. 7 a) a case of duplicity of an eye, and says that specimens occur in which each eye is doubled, so that there are two eyes at the base of each arm instead of one.

635. Stomobrachium octocostatum ( JEquoridse) : variety found in Cromarty Firth, frds of size represented by FORBES (Monogr. Br. Naked-eyed Jfedusce) ; ovaries bluish instead of orange, and without denticulated margins. Tentacles arranged in double series, long and short alternating, while in the type the series is single. The number of large tentacles same as in type. Each smaller tentacle bears vesicular body at base, without pigment or visible contents. The same variety figured by EHRENBERG, Abh. Ak. £erl., 1835, Taf. vm. fig. 7. ROMANES, G. J., Jour. Linn. Soc. xn. p. 526. [Simultaneous Variation of the several segments.]

With Nos. 634 and 635 compare the fact that in. Tiarops polytn there are //»/•/, m//// as a specific character four diadems 1 >et ween each pair of radial tubes, making in all sixteen instead of •////(/, which is the usual number in the genus. ROMANES, G. J., /-. I. luii. Sue. Zool., xn. p. 525. • i ;:{(>. Aurelia aurita. This form exhibits an exceptional frequency «f Mi-ri>tic Variation. In the normal there are 16 radial canals, 4 oral Inb.-s, 4- generative organs and 8 lithocysts. The departures from this normal form have been described in detail by KHUK.\]:I i:<.' and by ROMANES*.

.Men-tir Variation in Aurelia may occur in two distinct ways, tir.-t in tindecree to which there is complete separation between tin ^.-ii'-rat ive sacs, and second in actual numerical change. In the commonest form of Aurelia there are four generative organs each met from its nri^lilxuns, but iu some specimens the generative epithelium is continuous all round the mouth, and there is then one continuous generative cliunilii-r. though opening by 4 openings as usual. (Such absence of complete sepa- I'ati'in lii-tween some of the generative organs is not rarely seen in cases of numerical Variation, v. infra.) Though the epithelium is then continuous it does not form ;i triiicircle, but is sacculated to form 4 (as normally) 3, 6, or some other number of incompletely separated parts. EHBENBERG (I.e., p. 22) saw a case in which there were 6 such sacculations, three on each side being united and having one generative pouch, luit cadi of these pouches opened by 3 openings. There was thus a bilateral -\ -1111111 -try, i -ai-h halt' containing three lobes of ovarian epithelium incompletely -i ]iarat'<l t'mni each other. Complete union of all the generative organs was very ran .

Tinspecimens differ greatly with regard to the degree to which the generative • pithelium i- 1'ol.lcd off, and in the shapes of the generative organs. Commonly tin (.'enerative epithelium is of a horse-shoe form, the two limbs of the horse-shoe i"t Hinting eac-h ntlier; but in some specimens the two limbs maybe to various approximated, so that each generative organ is kidney-shaped or even roughly circular. (Cases figured by EURENBERG, 1. c., PI. u.) [Here note the Simultaneous Variation of the single quadrants.]

< >f these the most striking and also the most frequent are variations consisting in a perfect and symmetrical change in the fundamental number of segments composing the disc. Normally there are four • |u:idrants (Fig. ll'S, 1). Varieties are found having only half the usual number of organs, the disc being made up of two halves, each containing one generative organ (Fig. 128, IV). Other symmetrical varieties having three, and six, as their fundamental numbers are shewn in Fig. l->. V, and II. These figures are from ROMANES. Symmetrical forms ha\ing live segments and eight segments are described and figured by KiiKKNHKiic. As to the comparative frequency of these forms facts are invcn below. In each of them all the parts normally proper to one

• (iiadrant are repeated in each segment of the disc, the number of parts lieing greater or less than the normal in correspondence with the fundamental numliiT of the specimen. Next, the number of certain organs may vary independently of other organs. For example as seen in Fig. 128, VI the radial canal FIG. 128. Diagrams of various forms of Aurelia aurita, slightly simplified from ROMANES. I. The normal. II. Symmetrical form with 6 radii. III. Two additional chief radial canals in opposite interradii (where manubrial lobes also were bifid) and substitution of two canals for one in another iuterradius. IV. Form with two generative organs. V. Form with three generative organs. VI. Symmetrical form in which the intergenital canals are all doubled, the others remaining single. VII. Apparently upper half-disc arranged as for a symmetry of four, lower half for a symmetry of six. VIII. One of the quadrants tripled (?). IX. Form resembling VI. except that in one quadrant the intergenital canal is not doubled. The descriptions are not altogether those of ROMANES.

normally lying in a plane between each pair of generative organs may in each quadrant be represented by two canals, and in correspondence with this change the number of marginal organs is proportionately changed in the quadrants affected. But besides these changes symmetrically carried out in each quadrant or in the whole disc, one or more quadrants or a half-disc may vary independently. For example Fig. 128, VII, shews a specimen in which the two upper quadrants are normal but the lower half-disc is primarily divided into three. (In the case figured the parts of the lower half-disc

are not quit.- accurately distributed). Similarly a particular quadrant may be represented liy two sets of parts or by three sets (Fig. 128, VIII), tinother three quadrant^ being normal or nearly so. I have seen a case also in \\hii-h tinchief .symmetry was arranged a. - .for //,/••, >eg nifiits (having .". oral lobes), but one of the three segments was imperfectly divided into two. In a case of G segments, 3 on one side may be large and the other 3 small, somewhat as in Fi^. 1 28, VIII, but the whole disc was not circular, the i.nliu, on the >ide of the large segments being the greater.

In the figures (after ROMANES) all the discs are represented as circle-,, but my own experience was that when there was not a truly 83 minei ri.-al dist ribiitiou of the generative organs the half quadrant or other segment j,, which the number of parts was greatest bulged outwards, thus exemplifying the general rule that when an organ divides the two resulting parts are together larger than the undivided organ. Besides those specified, there are also irregular cases, e.g., specimens with .". generative organs but 4 oral lobes and other parts in multiples of I, but as Kiii:i.\i;i;i:<; says in such cases it is generally possible to detect that one i,f the generative organs is larger than the others or even partially double. He also saw cases otherwise arranged in a symmetry of G, but having 22 chief radial canals instead of 24, «fec. Also 14 radial i-anals (instead of 12; \\ere found in some cases of 3 generative organs.

As everyone will admit, it is impossible in regular threes, sixes, Arc. to say that any particular segment is missing or is added rather than another. *637. Among thousands of individuals seen by EHRENBEHG only two were S-rayed, \7i — 20 were Grayed, some 20 — 30 were 5- and 3-rayed, the remainder being I -rayed. In percentages, 90 are 4-rayed, 3 are 3-rayed 3 are 5-rayed, 2 are G-rayed and 2 have other numbers. Tinresult of an attempt to ascertain these percentages in a great shoal of A a n I'm washed ashore on the Northumberland coast on 4 Sept. 1892 is given below. The radial canals were not counted, and the numbers apply strictly to the generative sacs only. It will be seen that the proportion of abnormals is lower than that given by Ehrenberg.

There were therefore 1735 normals, 19 symmetrical varieties and 9 irregulars. It will be noted not only that the symmetrical varieties are comparatively frequent, but also that the several forms of irregularity were seen for the most part in single specimens only. The number of jaws in the pedicellariae differs in different forms of Echinoderms, and I am indebted to Professor C. STEWART for information concerning them. In Asteroidea the number of jaws is usually two, but in Luidia savir/nii the normal number of jaws is three.

In the Echinoidea the number of jaws is usually three, but in Asthenosoma the normal number is four. 638. Dorocidaris papillata : number of jaws in pedicellarise II. A pedicellaria with four jaws from the abactinal region. normally three as in Fig. 129, I, but occasionally four in pedicellariae of the abactinal region, as in Fig. 129, II. [Note that the variety is perfect and symmetrical] For this fact I am obliged to Professor Stewart, who kindly allowed this figure to be made from his preparations.

639. Luidia ciliaris : pedicellariae nearly all with three jams; but on Roscoff specimens a few having two jaws occur on the borders of the ambulacral groove. In BanyuTs specimens none such were found in this position, but there is one in almost all the marginal intervals. CUENOT, Arch. zool. exp., S. 2. V. bis, p. 18. 640. Asterias glacialis : occasionally threejawed pedicellarue like those of Luidia are found among the normal two-jawed pedicellarise. CUENOT, 1. c., p. 23.

*G4I. It was purposed at this point to have introduced an account of Meri>tic \ariations observed in the manner of division of nuclei and cells; }>\\\ I have found that, to give adequate representation of these tacts even in outline, it would be necessary not only to treat nt a very complex subject with which I have no proper acquaintance, but also greatly to enlarge the scope of this work. I'.ut were no word >aid on these matters, indications most useful ,i- ( -011111 K ait on the nature of Meristic Variation at large would have to be foregone; and unwilling that these should be wholly lost I shall venture briefly to allude to so much of the matter as is needful to shew some ways in which the facts of abnormal celldivision can be iix-d in reference to the wider question of Meristic Variation.

\\V have been dealing with cases of Radial repetition, and we ha\e seen that with Variation in the number of parts the result may still lie radially symmetrical. It therefore becomes of interest to note that in the case of abnormal cell-division the result of numerical change may in like manner be radially symmetrical. ('i 11s which should normally contain two centrosomes and which should divide into two parts have been seen to contain three centro-

Fio. 130. Triastors. I. Tripolar division of nucleus in embryonic tissue of Trout (after II I;NM '. rv1). II. Triaster from mammary carcinoma. Centrosomes not shewn (from FI.KMMINO-). somes (Fig. 180) prior to division into three parts, and the triangle formed by the three centrosomes maybe equiangular just as may be the triangle of the segments of the abnormal Aurelia (Fig. Ills, V). or of the jaws of the normal pedicellaria of Dorocidcvris (Fig. l^'.h. It is, I imagine, difficult to suppose that the radial symmetry of each of these series of organs is

different in its nature, or indeed that it is anything but a visible expression of the equality of the strains tending to part each segment from its neighbours. (The case of the triaster is taken as the simplest and most plainly symmetrical, but examples of cells with greater numbers of centrosomes, sometimes dividing symmetrically, have also been seen.) For our purpose this fact is first of use as a demonstration of the absurdity of an appeal to " Reversion " as a mode of escape from the admission that variations in Radial Symmetry may be total and perfect though the new number of segments is one which presumably never occurred in the phylogeny ; for we need scarcely expect that even conspicuous defenders of the doctrine that all perfection must have been continuously evolved, will plead that the cells of every tissue in which a triaster is found did once normally divide with three poles. Yet if it be once granted that the symmetry of these abnormal forms is a sudden and new departure from the normal, it will not be easy to put the other cases on a different footing.

Though we have repudiated all concern with the causes of abnormality, mention may be made of the fact that niultipolar figures, both regular and irregular, have been observed to result from the action of reagents (e.g. quinine, HERTWIG'). .Such figures are of course well known especially in the case of carcinomatous growths, and as Hertwig observes, from the resemblance of these figures to those artificially induced by chemical means it seems possible that these pathological appearances may also be the result of some chemical stimulus. But whatever be the immediate or directing causes of abnormalities in cell-division, or of those other abnormalities in the segmentation of Radial Series of larger parts, and whether any of the causes in the several cases be similar or different, we can scarcely avoid recognition that the resulting phenomena are closely alike2.

- See also a case of the presence of triasters in two bilaterally symmetrical tracts of the blastoderm of Lolicjo (v. infra). As seen in the majority of adult Echinoderms the repeated uiv arranged with a near approach to a Radial Symmetry .tin! it is thus convenient to consider their Meristic Variations in that connexion. Hut it must of course always be remembered that in their development these repetitions are in origin really u Successive Series and not a Radial Series. The segments are not all identical (as, in appearance at least, they are in many Coelenterutes iVe. ), but are morphologically in Succession to each ..ilicr, though there may be little differentiation between them.

In the case then-lure of Variation in the number of segments, resulting in the production of a body not less symmetrical than the normal body, there must be in development a correlated Variation among the several members like that seen in so many cases <>f additions to the ends of Linear Series. This cireuinsiuiiee should be kept in view by those who seek in cases of numerical Variation, in Echinoderms to homologize -eparate segments of the variety with those of the type, hoping to be able to say that such a radius is added, or such other missing. As in other animals, this has been attempted in EchinodermSj and though I know well that in the complex subject of Kchinoderm murpholugy I can form no judgment, yet it is difficult to suppose that the same principles elsewhere perceived would not be found to hold good for Kchinodernis also.

All that is here proposed is to give abstracts of facts as to Variation in the numbers composing the Major Symmetries. 1 1 will of course be remembered that though the fundamental number in Echinoderms is most commonly rive, other numbers also occur us normals, (e.g. four in the fossil Tetracrinus, six in some Ophiurids, \re. Examples will be given of total change from tive to lour and to six, and so on. It is besides not a little interest ing that of the normally 4-rayed Tetracrinus both ."> -rayed and .".-rayed varieties should be known.

llesides the e.\a m pies of t ot u I Variation t here are a few cases of incomplete Variation in which there is a fair suggestion that a particular ray is reduced in size (Nos. 680 and 681, &c.). There are also two cases of imperfect division of a ray in an Echinid (Nos. 688, &c.), while in Asteroids &c. this condition is common. It is of importance to observe that just as in Linear Series abnormal divisions of members of the series are commonly transverse to the lines of Repetition, so in radial forms the divisions of rays are commonly radial.

The evidence is complicated by the fact that in many Echinoderms extensive regeneration can occur, and in some genera reproduction by division of the disc and subsequent regeneration is almost certainly a normal occurrence1. Nevertheless it cannot be doubted that the variation seen in Echini, in Asterina, in the discs and stems of Crinoids, &c., are truly congenital. Similarly, though in Asterias &c. reduction in the number of arms might otherwise be thought to be due to mutilation, it cannot be so in Echini &c.

*642. Cucumaria planci : among 150 half-grown specimens found at Naples five were 6-rayed. LUDWIG, H., Zool. Anz., 1886, ix., p. 472 [These specimens are described in detail.] To determine which is "the intercalated ray" the following ingenious reasoning is offered, and as a good practical illustration of the conception of the individuality of segments as applied to an Echinoderm we may well consider it. 1 It is likely that several of the Ophiurids and Asteroids which normally have more than 5 arms undergo such fission. LUTKEN ((Efvers. Uansk. vid. Selsk. Fiirh., 1872, pp. 108—158 : tr. Ann. and Mag. N. H., 1873, S. 4, xn. pp. 323 and 391) gave an account of this phenomenon. Ophiothcla isidicola (Formosa) generally has 6 arms, rarely equal, usually 3 large opposite to 3 small; specimens common with only 3 arms, with appearance as if corresponding half-disc cut off. There can be no doubt that the animal divides and that the other 3 arms are renewed. The same phenomenon has been seen in other small 6-armed Ophiurids, especially of genus Ophiactis, but Liitkeu never saw any trace of it in any normally 5-rayed species of the genus. There are indications that the division occurs once when the animal is very small and again when it is adult or nearly so. In Ophiocomu pumila the small specimens have 6 arms, while the adults have 5. Probably therefore division only occurs in the young, the last division being followed by the production of 1 or 2 arms instead of 2 or 3.

Division is probably not a usual occurrence even in Ophiurids having more than 5 arms. Ophiacantha anomala has normally 6 arms, and 0. vivipara has 7—8, but no such appearances are known in them. Similarly there is evidence [figs, given] that certain Asteroids having normally more than 5 arms viz. Asterias problema Stp. [ = Stichaster attains], A. tennitpina &c. undergo fission ; but there is no reason for believing that other many-armed Asteroids divide. The Solasters have many rays, Asterias polaris has 6, but no signs of division are seen in them.

An account is also given of the comet-like specimens of Ophidiaster crilrarhis, occasionally found, having one long arm, at the adoral end of which are present 4 or 5 arms as mere tubercles or as half-grown structures. This phenomenon is well known in Linckia multiflora, in which doubtless the separate arms may break off, each reproducing complete disc and arms. [See also as to Stichaster albulns, Asterina wega, &c., CUENOT, L., Arch. zool. exp., V. bis, 1879—90, p. 128; and as to Linckia, SARASIN, Ergeb. naturw. Forsch. auf Ceylon, 1888, i. Hft. 2.]

In the nonii.-il there are •"> radii and interradii, and 10 tentacles: in tinabnormal* there are 6 and 12 respectively. In half-grown normals the .". ambulacra of the ventral trivium have more tube-feet than the '1 ambulacra of the bivium ; also the pair of tentacles corre- >l"iii«liiii,r to the central radius of the trivium are smaller than the rest. In the aliiionnals :'< amliulacra have more tube-feet and are separated bv narrower intcrradii than the rest, and of them the central has the le:iM pair of tentacles : therefore these are the 3 radii of the ventral tiivium, and of them the central is the central of the normal. The structure of the calcareous ring bears out this correspondence. The central radius of 1 he \entral trivium is therefore not the intercalated radius.