192 Prof. Carl Glaus on the 



new special assumption appears arbitrary and unfounded. 

 From the two axioms follows the division of the Siphono- 

 phora into two primary divisions, which Hackel denominates 

 Siphonanthffi and Disconanthse, and which, according to their 

 origin, would be referred, the former to the Anthomedusas 

 and the latter to the Trachymeduste. The inadmissibility of 

 this dipliyletic derivation has already been shown by another 

 hand, and the contradictions have been indicated which would 

 result for the structure and development of the Velellce from 

 the association with octoradial Medusaj *. It is not only that 

 the stage of the radiate Disconula only follows upon a simply 

 constructed bilateral stage of development, rendering it pro- 

 bable that here this is preceded by a bilateral division like 

 that in Siphonanthan larvffi, but also the mode of origin of 

 the mantle, which is by no means to be referred directly to 

 the Medusan umbrella, as well as the development of an 

 abundant vascular net and powerful muscular layer on the 

 aboral surface, in contrast to the non-vascular and non-mus- 

 cular exumbrella of the Medusae, cannot be reconciled with 

 Plaeckel's views. 



Against the Medusa-theory, however, in whatever form or 

 modification it may be put forward, I have in my former 

 })aper urged another argument, which has been entirely ignored 

 by liaeckel. I remarked that " another consideration renders 

 it improbable that the sexual form of the Hydroid polyps in 

 its perfect form furnished the starting-point for the production 

 of the Siphonophora, seeing that its ontogenetic origin is pre- 

 ludicl hy Hydroid- stocks^ which consequently, even in a Me- 

 dusa^ altered hy dislocation of 'particular parts of the body and 

 transformed into the stem form of the Siphonophora^ must have 

 recurred in the development of the latter ^ " The direct develop^ 

 ment [loithout alternation of generations) of individual Hydroid 

 Medusoi J is, however^ unquestionably only a sid)sequent secon- 

 dary condensation of the developmental process, which, there- 

 fore, we are not justified in taking as the starting-point of the 

 derivation^ The Medusa-theory, however, commences with 

 this subsequent, secondary, hypogenetic development of the 

 stem-form, which is already repeated as a Medusa in the 

 bilateral (Siphonula) or radial (Disconula) Siplionophorau 

 larva, and consequently leaves the older and originally meta- 



* Chun, Sitzimgsb. der k. preuss. Akad. der Wiss. Berlin, 1888, 

 Ed. xliv. pp. 3, 4. See ' Annals,' ser. 6, vol. iii. pp. 216-218. 



+ As supposed by Metschnikoff and also by Haeckel in his " Proto- 

 meda." 



\ To these bt'l(jug the Trach_vuiedus;e and aUo, therefore, Ilaeckel's 

 " ^ii'chiuieda,'" 



