August 26, 1897] 



NATURE 



407 



propagate by seeds. If the budded forms are sexual, while the 

 budding forms are not, there is an obvious explanation of the 

 difference in form. Even where there is no such fundamental 

 difference in function, the circumstances of early life are very 

 different, and may well produce an unlikeness upon which 

 Natural Selection may found a division of labour. 



No one who tries to trace origins can rest satisfied with 

 Steenstrup's account of alternation of generations. He makes 

 no effort to show how it came about. Instead of considering 

 alternation of generations as a peculiar case of development 

 with metamorphosis, complicated by asexual reproduction,' he 

 considers asexual reproduction as a peculiar case of alternation 

 of generations.- He ignores all the facts which show that the 

 alternation niay have been gradually attained, an omission 

 which is only excusable when we note that his treatise is dated 

 1842. He asserts dogmatically that there is no transition from 

 metamorphosis to alternation of generations. 



It is impossible to think much on this subject without falling 

 into difficulties over the word generation. For my own part I 

 believe that such y/oxAs!L% generation, individual, organ, larva, 

 adult cannot be used quite consistently in dealing with a long 

 series of animals whose life-histories vary gradually and with- 

 out end. Ordinary language, which was devised to meet the 

 familiar and comparatively simple course of development of man 

 and the domestic animals, is not always appropriate to lower 

 forms, with complex and unusual histories. If we are resolved 

 at all hazards to make our language precise and uniform, we 

 either fall into contradictions, or else use words in unnatural 

 senses. 



Certain recent discussions render it necessary to point out 

 that there can be no alternation of generations without increase 

 by budding. If a single larva produces a single sexual animal, 

 as when a pluteus changes to an Echinus, there is development 

 with transformation, but not alternation of generations. 



It is, I think, of importance to be able to resolve so peculiar 

 a phenomenon as alternation of generations into processes which 

 are known to occur separately, and which may have arisen 

 imperceptibly, becoming gradually emphasised by the steady 

 action of the conditions of life. Every startling novelty that 

 can thus be explained extends the application of that principle 

 which underlies the theory of Natural Selection — I mean the 

 principle that a small force acting steadily through a long time 

 may produce changes of almost any magnitude. 



The Hydrozoa yield innumerable and varied examples of 

 development with transformation and also of budding. They 

 yield also the most admirable examples of division of labour. 

 We have Hydrozoan colonies, such as a budding Hydra, in 

 which all the members are pretty much alike, but we soon 

 advance to differentiation of the feeding and the reproductive 

 members. In the Siphonophora the colony becomes pelagic, and 

 floats at the surface of the sea. Then the medusae no longer 

 break off and swim away, but are harnessed to the colony; and 

 drag it along. The colony may contain feeding polyps, which 

 procure and digest food for the rest ; swimming bells, which are 

 attached medusae ; perhaps a float, which is a peculiar kind of 

 swimming bell ; defensive polyps (which may be either batteries 

 of nettling cells or covering organs) ; and reproductive indi- 

 viduals. As the individuals become subordinated to the colony, 

 and lose essential parts of the primitive structure, they pass 

 insensibly into organs. 



The life-histories of Invertebrates abound in complications and 

 paradoxes. Thus Eucharis, one of the Ctenophors, becomes 

 sexually mature as a larva, but only in warm weather. This 

 happens just after hatching, when the animal is of microscopic 

 size. Then the sexual organs degenerate, the larva, which has 

 already reproduced its kind, grows to full size, undergoes trans- 

 formation, and at length becomes sexually mature a second 

 time.=* There is often a striking difference between the early 

 stages of animals which are closely related, or a strong adaptive 

 resemblance between animals which are of very remote blood- 

 relationship. In the Hydrozoa similar polyps may produce very 

 different medusre, and dissimilar polyps meduste that can hardly 

 be distinguished. There are insects so like in their adult state 

 that they can only be distinguished by minute characters, such 

 as the form and arrangement of the hairs on the legs, and yet the 



1 This is a convenient short account of Alternation of Generations, but 

 it will not apply to every case. In Hydra, for instance, there is an ill- 

 tlefined alternation of generations, but no metamorphosis. 



- C/. Leuckart, loc. cii., p. 183. 



3 Chun, Die pelagischt Thieiwelt, p. 62(1887). 



NO. 1452, VOL. 50] 



larvne may be conspicuously different. ^ Annelids and Echino- 

 derms yield fresh examples of the same thing. In Lepidoptera 

 and Saw-flies the larvie are very similar, but the winged insects 

 quite different.^ New stages may be added in one species, 

 while closely allied species remain unaffected. In Cunina and 

 the Diphyidre we get combinations which strain the inventive 

 powers of naturalists even to name. Natural Selection seems to 

 act upon the various stages of certain life-histories almost as it 

 acts upon species. 



But the history is not always one of growing complexity. 

 Sometimes, for example, a well-established medusa-stage is 

 dropped. First it ceases to free itself, then the tentacles and 

 marginal sense-organs disappear, then the mouth closes. In the 

 fresh-water Cordylophora the medusa is replaced by a stalked 

 sac filled with reproductive elements or embryos. The Lucer- 

 nariae present a single stage which seems to be polyp and 

 medusa in one. Hydra has no medusa. It is not always clear 

 whether such Hydrozoa as these are primitive or reduced. Even 

 the hydroid polyp, the central stage in the normal Hydrozoan 

 life-history, may be suppressed, and certain medusa in both of 

 the chief groups develop direct from the egg or planula 

 (Pelagia, Geryonia, .Egina, Oceania). There is no stage com- 

 mon to all Hydrozoa except the egg. The same thing may be 

 said of the Tunicates. 



The life-history of many Arthropods is to all appearance 

 quite simple. There emerges from the egg a spider, scorpion, 

 or centipede (in most Chilopoda) which merely grows bigger and 

 bigger till it is adult. But if, as in most Crustacea, the cir- 

 cumstances of the species call for a migratory stage, such a stage 

 will be added. In certain Decapod Crustacea (Penreus, Leucifer) 

 a nauplius and as many as five other stages may intervene before 

 the final or adult stage. Some of these larval stages are com- 

 mon to a great many Crustacea, but none, as we now think, 

 belong to the original phylogeny. If a resting or a winged stage 

 is wanted, it is supplied just as easily ; witness theholometabolic 

 insects. Here again, so far as we know, there is nothing abso- 

 lutely new.* The stages which seem new are merely exaggera- 

 tions for special purposes of sections of the life-history, which 

 were originally marked out by nothing more important than a 

 change of skin and a swelling out of the body. Let us not 

 suppose for a moment that it is a law of insect-development that 

 there should be larva, pupa, and imago, or that it is a law of 

 Crustacean development that there should be six distinct stages 

 between the egg and the adult. Any of these stages may be 

 dropped, if it proves useless — either totally suppressed, or tele- 

 scoped, so to speak, into the embryonic development. Lost 

 stages are indicated by the embryonic moults of some centipedes 

 and spiders, Limulus, many Crustacea, and Podura. The par- 

 thenogenetic reproduction of some immature insects, such as 

 Miastor, shows a tendency to suppress later stages. Perhaps 

 the wingless Thysanura are additional examples, but here, as in 

 the case of Hydra and Lucernaria, we do not certainly know 

 whether they are primitive or reduced. It seems to be easy to 

 add new stages, when circumstances (and especially parasitism) 

 call for them. Meloe, Sitaris, and Epicauta are well-known 

 examples. In some Ephemeridse the moults, which are poten- 

 tial stages, become very numerous, but as a curious exception to 

 a very general rule, the last moult of all, which is usually so 

 important, may be practically suppressed. The fly of an 

 Ephemera may mate, lay eggs, and die, while still enveloped in 

 its last larval skin. 



Among the many cases of what one is inclined to call rapid 

 adaptation to circumstances (the chief indications of rapidity 

 being the very partial and isolated occurrence of remarkable 

 adaptive characters) are those which Giard ' has collected and 

 compared, and which he refers to a process called by him 

 Poecilogony. A number of very different animals * produce 

 according to habitat, or season, or some other condition closely 

 related to nutrition, eggs of more than one sort, which differ in 

 the quantity of nourishment which they contain and in the de- 

 gree of transformation which the issuing larva is destined to 

 undergo. The analogy with the summer and winter eggs of 



1 Some species of Chironomus are referred to. 



2 Baron Osten Sacken (Bert. Entoin. Zcits., Bd. xxxvii. p. 465) gives two 

 cases of Diptera, in which " almost similar larva: produce imagos belor.ging 

 to different families." 



*" Nirgends ist Neubildung, sondern nur Umbiidung." — Baer. 

 *C. R. 1891, 1892. 



S E.g. Crustacea (Palaemonetes, Alpheus), Insects {Musca corvina, some 

 Lepidoptera and Diptera), an Ophiurid (Ophiothrix), a Compound Ascidian 

 Leptoclinus), &c 



