Secondly, as regards the last-mentioned author's explanation of such serial homology as exists in the centipede and its allies, the very groundwork is open to objection. Multiplication by spontaneous fission seems from some recent researches to be much less frequent than has been supposed, and more evidence is required as to the fact of the habitual propagation of any planariæ in this fashion.[[169]] But even if this were as asserted,
nevertheless it fails to explain the peculiar condition presented by Syllis and some other annelids, where a new head is formed at intervals in certain segments of the body. Here there is evidently an innate tendency to the development at intervals of a complex whole. It is not the budding out or spontaneous fission of certain segments, but the transformation in a definite and very peculiar manner of parts which already exist into other and more complex parts. Again, the processes of development presented by some of these creatures do not by any means point
to an origin through the linear coalescence of primitively distinct animals by means of imperfect segmentation. Thus in certain Diptera (two winged flies) the legs, wings, eyes, &c., are derived from masses of formative tissue (termed imaginal disks), which by their mutual approximation together build up parts of the head and body,[[170]] recalling to mind the development of Echinoderms.
Again, Nicholas Wagner found in certain other Diptera, the Hessian flies, that the larva gives rise to secondary larvæ within it, which develop and burst the body of the primary larva. The secondary larvæ give rise, similarly, to another set within them, and these again to another[[171]] set.
Again, the fact that in Tænia echinococcus one egg produces numerous individuals, tends to invalidate the argument that the increase of segments during development is a relic of specific genesis.
Mr. H. Spencer seems to deny serial homology to the mollusca, but it is difficult to see why the shell segments of chiton are not such homologues because the segmentation is superficial. Similarly the external processes of eolis, doris, &c., are good examples of serial homology, as also are plainly the successive chambers of the orthoceratidæ. Nor are parts of a series less serial, because arranged spirally, as in most gasteropods. Mr. Spencer observes of the molluscous as of the vertebrate animal, "You cannot cut it into transverse slices, each of which contains a digestive organ, a respiratory organ, a reproductive organ, &c."[[172]] But the same may be said of every single arthropod and annelid if it be meant that all these organs are not contained in every possible slice. While if it be meant that parts of all such organs are contained in certain slices, then some of the mollusca may also be included.
Another objection to Mr. Spencer's speculation is derived from considerations which have already been stated, as to past
time. For if the annulose animals have been formed by aggregation, we ought to find this process much less perfect in the oldest form. But a complete development, such as already obtains in the lobster, &c., was reached by the Eurypterida and Trilobites of the palæozoic strata; and annelids, probably formed mainly like those of the present day, abounded during the deposition of the oldest fossiliferous rocks.
