74 JIMBRYOLOGT 



phora, so tliat, iu pursuance of the medusa theory, we are 

 required to distinguish iu the Siphonophora two highly 

 heteromorphous generations, the first, produced from the 

 egcr, consti'ucted upon the Siphonula type, and reproducing 

 by budding only, the second a generation of fertile indi- 

 viduals not bilaterally symmetrical, and without dislocation 

 of the primary organ. Still sharper, perhaps, is the contrast 

 between the Disconula of the Vellelidse, which is i-eferred 

 by Haeckel to certain Trachomedusae, and the structure of 

 the Chrysomitras. 



On the other hand, for the hydroid theory there is the 

 difficulty of explaining how a firmly attached hydroid stocklet 

 could detach itself and become metamorphosed into a free- 

 moving, pelagic organism. If, however, "we assume that a 

 hydroid stocklet attached itself by means of a broadened 

 basal plate to the surface of the water, instead of to a fixed 

 body, as may occasionally be observed in the Scyphistomas, 

 and acquired under favourable circumstances the power to live 

 on in this condition, then with this conception the transition 

 from the attached to the free mode of life is brought about 

 by a floating at the surface of the water, a form of locomotion 

 which has been retained in Physalia and Vellela. Nay, we 

 need only to conceive that the flattened basal part of the 

 stem, which attached itself to the under-surface of the watei-, 

 curved inwards like a canoe, and finally became, with its peri- 

 sarc-covered face, completely invaginated,^ in order thus to 

 make the phylogenetic origin of the pneumatophore conceiv- 

 able, and to support this conception by the consideration 

 that such a course of development must have been constantly 

 accompanied by certain advantages to the entire colony. 

 Not until after the development of this hydrostatic apparatus 

 would a separation from the surface of the water and a 

 descent into greater depths become possible. The pneuma- 

 tophore would accordingly be the first, most primitive oi'gan 

 by the development of which the characteristic peculiarities 



' It has actually been observed in the planula of various Cnidaria that 

 the future jjoint of attaclunent, which lias undergone glandular alteration, 

 is more or less invaginated, as in the Scyphomedusie and in Eutima 

 (Brooks). 



