880 SYSTEMATIC HISTOET OF THE INEUSOEIA. 



striated, with the habit of Cocconema, | middle and indistinct striae. Rab D. 



but suddenly reflexed under the very j p. 26, pi. 10. f. 1. Salzburg. Frustules 



obtuse apices in the manner of Eunotia. ! minute. 



= Glceonema Sinense, ERBA. 1847, I t\ t ./• 7 cv • 



p. 484; EM. China, Java. Ehrenberg-'s I ^ Doubtful Species. 



figure represents the valve distinctly I E. globiferiim (Ag.). — Filaments ab- 



striated, with straight venter, very ob- | breviated, frustules simple or binately 



tuse, rounded ends, and the dorsal mar- [ conjoined, hyaline, with a globule in the 



gin very convex, and curved upwards at middle. = GlcBonema glohiferum, Ag CD. 



each end. | p. 31. Italy. 



E. gracile (Rab.). — Valves slender,! 'E. Arcus= Glceonetna Arcus, ERBA. 

 with truncate apices^ slightly gibbous ; 1856, p. 333, f. 26. Africa. 



Genus AMPHORA (Ehr.). — Frustules free, cymbiform ; lateral view lunate 

 or arcuate, with a nodule at the middle of the ventral margin ; front view 

 with the median lines and nodules of valves approximate and within the 

 margin. The frustules are mostly very thin, hyaline, and imperfectly sili- 

 cious: their form is peculiar; and Professor Arnott, who has given in the 

 sixth volume of the 'Microscopic Journal' a detailed account of their structure, 

 aptly compares it to that of " a coffee-bean rounded on the back and hollowed 

 out in front." Many of the species are insufficiently known ; they should be 

 viewed in front, back, and side. Fortunately, from their hyaline natui'e, the 

 dorsum and venter can in most cases be examined merely by the alteration of 

 the focus. The lateral view closely resembles that of a Cymbella, but has the 

 nodule marginal. The front view is usually barrel- shaped, owing to the con- 

 vexity of the valves, which are so curved inwards that their central nodules 

 are more or less approximate and frequently appear nearer to the connecting 

 zone than to the margins. The portions of the valves interior to the median 

 line are inconspicuous, and rarely afford diagnostic aid ; whilst the portions 

 exterior to the median lines are important, offer the best view of the trans- 

 verse striae, and vary in shape according as the median line appears straight, 

 concave, or flexed. In our descriptions we call these latter the outer por- 

 tions, and when they project inwards at the centre in a cuneate manner, or 

 appear inflexed, we term them canoe-shaped. The dorsum is convex, shows 

 no nodules or lateral valves, and is mostly marked by longitudinal lines 

 between longitudinal series of short, transverse striae, like the connecting 

 zone of StriateUa, but unaccompanied by internal plates. The late Pro- 

 fessor Gregory, who directed attention to these facts, believed that Amphora 

 could be divided into two groups — simjole and complex, from the absence or 

 presence of this structure. His arrangement, however, we are unable to 

 adopt, because in many species a decision is difficult ; and indeed we think 

 that the longitudinal lines, so common if not invariable, indicate the complex 

 structure, although the hyaHne natiu'e of the frustules may interfere with its 

 detection. In ximphora the specific characters are taken, almost constantly, 

 from the front \dew, and not from the lateral one as in most other genera of 

 Diatomaceae ; and as the connecting zone varies greatly in breadth according 

 to the condition of the frustule, due allowance must be made for that fact. If 

 division has recently taken place, the connecting zone will be narrow, and the 

 ends of the frustule less broadly truncate than just previous to that process. 

 For the same reason we believe that the number of longitudinal lines varies 

 and affords no aid in distinguishing the species. Amphora contains several 

 species of Agardh's genus Cjonbella, and ought, in our opinion, to have re- 

 tained that generic appellation. Because of its cymbiform frustules, we have 

 removed this genus from the Naviculeae, where Kiitzing placed it ; the same 

 reason, added to the presence of median lines and nodules, compels us to 



