38 W. M. BALE ON SOME OF THE DISCOID DIATOMS. 



Some valves which I have seen with this character were in all 

 other respects similar to normal valves of A. undulatus, among 

 which they occurred, and I see nothing to justify their separation, 

 the slight divergence from the rectangular arrangement of the 

 striae being no more than is often found in P. formosum. Some- 

 times the striae meet at more than a right angle, so that the 

 third set is radial instead of tangential. If Schmidt's species 

 were accepted, this should make another species ! The striae are 

 sometimes nearly or quite obliterated on small patches at the 

 outer angles of the secondary areas, and occasionally along the 

 margins ; in some forms again they are wanting or represented 

 only by a few scattered puncta on a great part of those areas. 

 The umbilicus varies greatly in size, and may be either hexagonal 

 or may have three concave sides. Much variation exists in the 

 extent to which the areas are inflated, or, in other- words, in the 

 depth of the undulations. 



A consideration of the variations of this diatom will show how 

 many features there are which, met with in isolated forms, may 

 lead to the undue multiplication of species. 



In several species, perhaps in the genus generally, there is 

 a tendency to produce valves in which the secondary areas or 

 " Nebenfelder " are replaced by primary ones or " Hauptfelder," 

 so that all the areas become alike, except in their elevated or 

 depressed condition. Van Heurck has figured such a form of 

 A. undulatus the forma sexapjiendicidata, which he says may 

 co-exist in the same frustule with the normal form. He refers 

 only to the presence of a process on every area, and does not 

 mention that the areas are otherwise modified, which, however, 

 I have always found to be the case. Other varieties of 

 A. undulatus exhibit the same tendency ; thus the large forma 

 maxima found in the Nottingham deposit is accompanied by its 

 "forma sexappendiculata" as also is an equally large variety 

 which only differs from it in the strongly apiculate margin. In 

 all these cases the compartments all correspond exactly with the 

 normal primary areas, both in the striation and the coarser 

 secondary markings. There may possibly be varieties with this 

 as the usual condition, as I have found one or two such forms 

 sparsely distributed in material where I noticed no typical valves 

 to which they might correspond. 



In A. Heliopelta also valves are formed in which all the areas 

 are alike, instead of alternately primary and secondary. 



It is to be noted that in all species where this phenomenon 



