148 THE INVERTEBRATA 



might be supposed to belong to each other. In more advanced stages 

 of development the displacement becomes extreme and no corre- 

 spondence can be traced. Lastly, with the great development of the 

 gas-secreting pneumatophore, the medusa bell is suppressed. 



While the above description gives an impression of the order 

 regarded as colonial animals the siphonophores must be primarily 

 considered as coelenterates exhibiting growth variability to such an 

 extent that the identification of the component structures as organs 

 or individuals is difficult and of purely academic interest. 



Order GRAPTOLITOIDEA 



Extinct, probably planktonic, Hydrozoa in which the polyp genera- 

 tion is dominant; the medusoid generation possibly represented by 

 gonophores; the individuals are budded off from one another and 

 remain in contact with the parent ; there is no definite coenosarc ; and 

 the perisarc is produced round the polyps as hydrothecae. 



Graptolites are represented in the earliest fossiliferous rocks, the 

 Cambrian, Ordovician and Silurian. Though we know nothing of 

 their soft parts, the exoskeleton was horny or chitinous and so is well 

 preserved. It resembles in general development that of the colonies 

 of the Calyptoblastea, in that it was produced round the polyps to 

 form definite hydrothecae. The graptolites, however, diflfer from 

 calyptoblast hydroids because new individuals are budded off 

 directly from older ones and not from a common coenosarc. Each 

 colony originates from a conical body called a sicula^ the exoskeleton 

 of the first formed individual. From the side of this a bud is formed, 

 and from the second individual which thus develops a third, and so 

 on. In this way a linear series of polyps is produced which are 

 arranged in a slender lamella {stipe), the hydrothecae being in contact 

 and the cavity of the colony being continuous. The simplest arrange- 

 ment is in Monograptus (Fig. 123) in which the hydrothecae are all 

 on one side of the stipe. In Diplograptus budding takes place altern- 

 ately on the two sides of the terminal individuals so that there is a 

 biserial stipe, and in Didymograptus (Fig. 124) the second and third 

 hydrothecae go on budding independently, so that we have a colony 

 with two stipes or branches. By another modification later hydro- 

 thecae bud off two individuals instead of one, and colonies like 

 Tetragraptus (Fig. 125) and Bryograptus are formed. 



In the absence of a coenosarc the graptolites were not attached by 

 a creeping hydrorhiza, such as occurs in Calyptoblastea. There was, 

 however, a thread coming off from the end of the sicula which ended 

 in a disc, by which it is supposed that the graptolites were attached 

 to floating seaweed (Fig. 1 24 B) . It is also possible that some graptolites 



