470 R. KIDSTON AND D. T. GWYNNE-VAUGHAN ON 



Here, again, we are in line with Dr Bertrand, who uses exactly the same principles 

 in deriving the more typical Zygopterid traces from that of Clepsydvopsis antiqua, and 

 we had just reached this stage in our speculations when his stimulating monograph 

 appeared. His important and extensive researches have cleared away many former 

 misconceptions, and have brought to light a large number of new facts, and as a result 

 of his investigation he has formulated a very interesting theory of the origin of the 

 Zygopterid petiole. In accordance with this theory he has classified the various types 

 into five series, which he regards as indicating natural lines of descent, and he gives a 

 close and detailed account of the interrelationships of the several series. We quite agree 

 with Dr Bertrand in so far that there seem to be several different lines of development 

 in the group ; but so much information is still wanting, not only as regards detail but 

 in many cases even concerning the most critical points, that we fear the chances are 

 against the ultimate success of so precise and exact a classification as he has adventured. 

 Still, it is only by classifying the facts at our disposal that we can arrive at any sugges- 

 tion of the actual lines of descent that are believed to exist in the order. In view, 

 therefore, of the interesting results that Dr Bertrand has arrived at by the application 

 of his theory, we feel justified in following his example by grouping together some of 

 the best-known forms on the basis of the common possession of what appear to us to be 

 the more fundamental characters. We must acknowledge our indebtedness to Dr 

 Bertrand for many of his facts and some even of his conclusions, which, somewhat 

 ungratefully, we are about to use in support of a rival theory. Of course, the practical 

 utility of any such grouping must depend entirely upon the real fundamental import- 

 ance of the characters that are selected as criteria. This in turn, in the absence of 

 reliable information upon so many points, depends so much upon the personal factor 

 that it is obvious that our arrangement may be seriously upset at any time by new- 

 discoveries. 



The most important and far-reaching distinction to be found among the different 

 Zygopterid petioles appears to us to be the presence of a single row of appendages 

 ("pinnae") on each side of the main rachis (text figs. 7-9), or of two such rows 

 (text figs. 10-14). This distinction seems also to be correlated with two special types 

 of " pinna " trace. In the first case the xylem strand of the pinna at the point of its 

 departure from that of the main rachis has always the form of a closed ring (text figs. 

 7-9). In the second it has the form of a slightly curved band which ultimately divides 

 into halves (text figs. 10 and 1 1). In some cases, however, the band is represented by 

 two separate halves even from the first, and the island is converted to an open hay 

 (text figs. 12 and 13). 



Another distinction of minor importance is the degree of extension of the lateral 

 islands of parenchyma, relative to the arms of the H. In some cases the island never 

 reaches far beyond the lateral termination of the cross-bar of the H (text figs. 1 2 and 13). 

 In others it reaches right up to the ends of the arms (text figs. 8-1 1). 



In applying the first and most important of these criteria it is found that the 



