308 ' ENDEAVOUR " SCIENTIFIC RESULTS. 



Nutting has noticed that among the corbula-bearing species, 

 when the proximal part of the gonocladium bears one hydro- 

 theca only, that one is generally normal, but that where 

 there are several hydrothecse in that position they are usually 

 more or less modified ; also that in Aglaophenia the former 

 type prevails, in Thecocarpus the latter. I find that the 

 same rule applies generally (though not universally) to the 

 Australian species. In A. megalocarpa, A. armata, and A. 

 tenuissima there are below the corbula several hydrothecse, 

 which are modified to a greater or less extent ; in A. caly- 

 cifera the hydrothecse themselves are little if at all altered, 

 but the internodes supporting them differ from the normal 

 hydrothecal internodes in being longer, and in the absence of 

 septal ridges. In A. brevirostris, on the other hand, there is 

 only a single hydrotheca, not distinctly modified, so that the 

 condition is the same as in Aglaophenia proper ; it is note- 

 worthy, however, that in this species there are generally one 

 or two modified hydrothecse on the distal part of the gono- 

 cladium, which projects beyond the corbula. 



In the crucialis group there are no hydrotheca? at all on the 

 gonocladium, the proximal internodes supporting only a row 

 of large sarcothecse, like the proximal part of a branch, and 

 the same condition is found in A. phyllocarpa and A. brachiata. 

 None of the numerous American species comprised in 

 Nutting's monograph exhibit this character, which appears 

 to be confined to typical Aglaophenice, unless A. brachiata is 

 3,n exception. 



Species which are known to belong to the divaricata group 

 are A. divaricata (Busk), A. acanthocarpa, Allman, A. laxa, 

 Allman, A. plumosa, Bale, and A. whitdeggei, Bale. The 

 same type prevails throughout all these species. The 

 corbula-ribs are narrow, armed on each side with sarcothecse, 

 which are longer than those usually found in the other 

 groups, the lowest two sarcothecse on the distal side are 

 without corresponding ones on the other side, and of these 

 two sarcothecse the first one is usually bifid or double, a 

 feature which Allman accounts for by supposing it to repre- 

 sent the lateral pair of the suppressed hydrotheca. Of the 

 species mentioned I have only observed the gonosome in 

 A. divaricata (including A. mccoyi), A. laxa, and A. plumosa, 

 but Allman's figure of A. acanthocarpa and Stechow's of A. 

 whiteleggei show the same structure, and though Hilgendorf's 

 figure of A. laxa does not exhibit similar regularity, I find it 

 in the specimen which I have examined. The hiatus caused 

 by the absence of the two lowest sarcothecse on the proximal 

 .side corresponds to the gap which in the crucialis group and 



