950 MR. H. 11. BKiyBLEY o>' EEPHODL'CED [Dec. 13, 



instances is that afforded by the scaling of the reproduced tail in 

 certain Lizards (12, 13). In a recent note Giard (36) has collected 

 instances of this kind of variation under the title of " hypotypic 

 regeneration," and has included thereunder the 4-jointed repro- 

 duced tarsus of Blattidae and Phasmidae. lu the suggestion that 

 such structures are really reversions to ancestral forms he is followed 

 by Bordage (6). 



Such a view must be based on the assumption that the normally 

 4-jointed tarsus of Lepismidte and Locustida; represents the primi- 

 tive condition in Insecta, the grounds for which being that the 

 Lepismidfe are usually held to be primitive forms and that the 

 rei^roduced tarsus in Locustidfe does not exhibit any reduction in 

 the number of its joints. But the tarsus of Insects as known to 

 us is characteristically _^w-jointed, and our ignorance of the mean- 

 ing of the phenomena known as reversions denies much weight to 

 arguments supported by appeal to them. 



Our present knowledge of the whole subject of reproduction 

 after injury is so scant}- as to render of very minor value such 

 arguments as that " it is advantageous for a mutilated individual 

 to abridge the process of reproduction and not to recapitulate in 

 their entirety all the phylogenetically ancestral stages." Have we 

 indeed any justification at all for supposing that reproduction of 

 any part is a recapitulation of even the normal ontogeny ? In 

 the cases already described only one seems to afford any degree of 

 suggestion that reproduction involves a throwing back of normal 

 development, and that is the observation of Blackwall that male 

 Spiders do not develop the adult sperm-case of the terminal joint 

 of the pedipalp when that appendage has been mutilated between 

 the penultimate and final ecdyses. But this peculiarly adult 

 structure is not a distinct joint, and the instance is one in which 

 a certain identihable part of the normal appendage is absent, 

 and so is unlike the instances in which the reproduced gro\\th is a 

 completely functional structure but differs from the normal in the 

 arrangement of its parts generally. 



It is of much interest in connection with the peculiarities of re- 

 production forming the subject of this paper, that departures from 

 the normal in the mam similar to them have been observed in 

 genera the nature of whose developmental history, and the fact 

 that the abnormal condition was frequently manifest symmetrically 

 on the two sides, seem to render it most unlikely that reproduc- 

 tion had occurred. Cases of this kind in antennas of certain 

 Hymenoptera and Coleoptera have been commented on by Bateson 

 (2. p. 411) ; and though m some of the examples the variation in 

 the number of anteniial joints was so great that the normal number 

 remained uncertain, the following features occurred not infre- 

 quently. "Where the terminal joint or joints were in the normal of 

 specialized structure, the exceptional cases of few-jointed antennae 

 presented a similar condition. Moreover there were usually de- 

 partures from the normal m the relative lengths of the other joints, 

 and the joints were usually of longer individual lengths than those 



