98 GENEKAL HISTORY OF THE INFUSOEIA. 



T. favus in the centre, we may diverge in lines to a circumference ending in 

 one line, in the long-anned T. Solennoceros, itself nearly resembhng Desmi- 

 clium tridens or D. hexaceros ; in another line ending in a form resembling 

 Desmldium apiculosum ; in another like Zygoceros rhombus, especially in the 

 front view ; in another analogous to Amphitetras antediliwicma ; and in 

 another to Campylodiscus crihrosus.^' 



Next after size and form, markings existing on the surface or within the 

 frustules have been employed as specific and generic characteristics ; but with 

 these, as with the fonner conditions, great uncertainty prevails in their ap- 

 plication, as we have ah-eady seen in the difference of opiuion, regarding some 

 internal markings of Grammatophora, between Dr. Gre\alle and Prof. Smith. 

 In like manner the character, the breadth, the relative position and distribu- 

 tion, the distinctness and the number of striae on the valves, although 

 tolerably constant in some species, are, in the majority, subject to gi-eat 

 variation. Then again some natui'alists coimt the number of striae in a 

 given space, as, for example, in the yxKui^^ ^^ ^^ inch, whilst others advo- 

 cate counting the entire number in the length of the valve. The latter plan, 

 to all appearance, must afford more certainty, although the trouble of it is 

 much greater ; for in the growth of fmstules there would seem an expansion 

 of theii^ walls, inducing consequently a displacement of the striae further 

 apart ; and observation does not confii^m the opinion, that in the imperfectly 

 developed frustules a smaller number exists, which are added to in course of 

 growth. 



However, just as in the case of the form and size, so, in this matter of the 

 superficial markings, there -udll be variations according as the frustules result 

 from self- division and are stereotyped impressions of an already existing 

 form, or according as they originate from sporangial frustules and may have 

 an individual idiosyncrasy, or be modified in their development by the locality, 

 and by surroimding cii'cumstances, season and the like. 



A writer in the Mic. Journ. (1855, p. 309) invites notice to another cir- 

 cumstance : — " Sufficient attention has not yet been paid to the sporangial 

 state of the Diatoms. Prom the observations recorded by Thwaites, Smith, 

 and others, different genera seem to follow different laws on the subject. In 

 Navicida this state appears to be always accompanied by a great dilatation of 

 the frustule, and the formation of a strong line or band between the median 

 line and the margin ; sometimes the new line is nearly straight and parallel 

 to the median Hue, except near the nodule, Avith which it seems connected ; 

 sometimes it is curved ; but whether both structures occur in the same 

 species, or are indicative of different species, no evidence has hitherto been 

 addaced The striae appear, however, to preserve nearly the same in- 

 clination to the new or intermediate Hues which they did in the non- 

 sporangial state to the median line ; and hence the direction of the striae is 

 not sufficient of itself to distinguish species, however good a character it may 

 afford, unless regard be had to the peculiar state of the fi-ustule." 



Prof. Smith has endeavoui'ed to frame some general rules for the guidance 

 of naturalists in instituting generic and specific characters, which we cannot 

 do better than subjoin in an abridged form {J. M. S. 1855, pp. 132-134; and 

 Synops. vol ii. p. xxii). In determining specific character, three circumstances 

 are of essential importance: 1. the structm^e of the valve ; 2. the habitat; 

 3. the arrangement of endochome in the li\ing frustule. 



The first can be applied to both living and dead or fossil specimens, and 

 affords the most constant and obvious characters. " These varieties of struc- 

 ture arise from the modes in which the silex combines with the, cellulose of 



