120 PRIMITIVE FORMS OF LIFE 



develop before the gastromerids. Nothing equals the brilliant 

 coloration and the richness and variety of form manifested 

 by these swimming hydrodemes, constituting the order of 

 Siphonophora. They are real autonomous organisms, and 

 furnish definite proof that what used to be called a colony 

 is nothing but the first phase in the formation of higher 

 organisms. 



The advance made by the medusas in the development of 

 Siphonophora can take place to an equal extent in the fixed 

 hydrodemes. By their help we can construct the stages right 

 up to that moment when the development of the egg ends, not 

 as in a hydrozoid which itself produces a hydrodeme, but 

 directly in a medusa. The gradual suppression of the 

 hydrodeme is here equivalent to the gradual suppression of the 

 prothallus in the Vascular Cryptogams. These medusae, in- 

 dependent of any hydrodeme, but whose formation was 

 prepared, and could not have been produced but for a 

 lengthy elaboration by a series of hydrodemes, themselves 

 undergo important modifications, becoming complicated in a 

 variety of ways till they finally attain many decimetres in 

 diameter. They form the class of Acalephae, comprising many 

 orders, 1 the first culmination of the Hydroid series. 



There is a second condition more important still, attained 

 by the Coral polyp via the Madreporaria, those very wonder- 

 ful reef -builders. Certain Hydroids akin to Hydractinia 

 had already been able to secrete lime. This property 

 is general among the hydrocorallines — so well studied by 

 Moseley in the course of the Challenger Expedition. In the case of 

 the hydrocorallines we can trace from the Echinopora, still closely 

 akin to Hydractinia, to Millepora, Allopora, Stylaster and 

 Cryptohelia, all the phases in the grouping of a certain number of 

 dactylomerids around a gastromerid, analogous to that which has 

 given rise to the medusae. But here the gastromerid, instead of 

 remaining independent in the centre of its ring of dactylomerids 

 attaches itself to them throughout its whole length and com- 

 municates with them by means of corresponding longitudinal 

 slits. It loses its tentacles, which are replaced by dactylomerids. 

 The whole forms a coralozooid achieved by a mechanism analogous 

 to that which produced the dialypetalous flowers with inferior 

 ovaries. 



1 XLIII, 640. 



