1896. THE STRUCTURE OF THE GRAPTOLITES. 191 



old called " the common canal," namely, the set of the inner ends of 

 the thecae, Tornquist names the " biserial chamber." If, as in 

 Climacograpius hickersianns, a longitudinal septum is present, the 

 " biserial chamber " passes at the beginning of the septum into two 

 •" uniserial canals." If, with Tornquist, we regard the peridermal 

 units as representatives of units in a once living colony, then the 

 sicula, the mother-animal of the colony, does not send out one 

 individual but a kind of budding stolon, which passes, in the following 

 order, through the "connecting canal," the "biserial chamber," and, 

 if there is a septum, the " uniserial canals.' In the " biserial 

 •chamber " this stolon buds alternately to either side ; in each of 

 the " uniserial canals " it buds to one side only. 



The family Dicranograptidae is still comparatively little known. 

 ■On the other hand, the families Dichograptid^e and Phyllograptidas 

 are better known, and this especially through Holm's examination 

 ■oi Didymogvapius, Teiragraptns, and Phyllogmpi us {16). It appears that 

 the first stages of development of the rhabdome are, in the main, 

 identical in these genera with the stages in Diplograptidae, wherefore 

 I need not give any further description of the structure of the sicula 

 end in these families. Holm, however, seems not to have decided 

 whether the individuals were formed by budding from a coenosarc 

 filling the "common canal," or from each other. He, therefore, some- 

 times speaks of the sicula in Tetvagraptns and PhyUogvaptns as giving 

 "origin to thecas, which as in Didymogvaptus and Diplograptiis are 

 developed from each other, and therefore occupy different heights 

 within the polypary," and sometimes of a " bud " and a " connecting 

 ■canal." By " bud " Holm means the same structure as Tornquist 

 calls " connecting canal," which I consider as nothing more than that 

 part of the first theca that lies between the passages to the sicula and 

 the second theca respectively {a on Fig. 3). By " connecting canal " 

 Holm means something quite different to what Tornquist does, 

 namely, the structure that I regard as the elder part of the second 

 theca. 



J. Hopkinson (6) has found specimens of Teiragraptus serra and 

 Didymogvaptus extensus in which the common canal is both divided by 

 transverse walls into as many chambers as there are accompanying 

 thecae, and partitioned off from the thecae by walls. 



Quite recently, at the end of 1895, H. A. Nicholson, and J. E. 

 Marr (19), taking the shape of the thecae as a marked character, 

 regularly inherited, have arranged a genealogical tree for the family 

 Dichograptidae, in the same way as in 1895 ^ derived different groups 

 •of Monogvaptiis, each by way of its corresponding Dimovphogyaptus, 

 from Diplograptns and Climacogvaptus respectively (18). 



Fig. 9, taken from Ruedeman (17) shows that colonies of 

 Diplograptiis were united by their virgulae into brush-like or almost 

 star-shaped colonies of a higher order, within which there was a 

 division of labour. Beside the Diplograpti in the ordinary sense, 



