1910.] NATURAL SCIENCES OF PHILADELPHIA. 461 



by various methods, and primarily by burning on the cover-glass and 

 mounting in Canada balsam. I owe to Mr. C. S. Boyer, who received 

 some of this material, the first intimation that these groups of four 

 frustules may come intact through a careful boiling in nitric acid and 

 bichromate, continued long enough to destroy entirely all accompany- 

 ing organic matter, and with it, of course, that belonging to the 

 diatoms themselves — a fact I have confirmed by repeated trials. Dis- 

 integration of groups takes place only on vigorous boiling and agitation. 

 This indicates the grouping to be maintained by silicious cementing 

 of the edges of the valves. The peculiar habit of growth seems per- 

 sistent, and this will appear more clearly in subsequent notes of the 

 occurrence of these groups at widely separated places. The species 

 may be described as follows: 

 Navioula socialis sp. nov. PI. XXXV, figs. 1 and 2. 



Frustules normally in motile groups of four, held with girdle-sides 

 together by a silicious cementing of adjacent valve-edges, and enclosed 

 in a common coleoderm. Valves linear-elliptic, sides slightly concave, 

 ends obtusely rounded. Transverse costae prominent, nearly parallel 

 except around terminal nodules, alike on both valves, 8 in 10 [x, wider 

 than intervening spaces, showing obscurely a longitudinal band. 

 Raphe nearly straight and simple. Longitudinal hyaline area nearly 

 one-third the width of valve, much contracted near terminal nodules, 

 slightly and unsymmetrically expanded around central n< dt le. Length 

 of valve, 60-120 //, breadth 13-27 /*. 



Fresh water: Swampy pools near Media, Pa. Type in cabinet 

 of T. C. Palmer. 



This species has a superficial resemblance to some forms of Navicula 

 viridis Kiitz. and the possibility is recognized that frustules isolated 

 from normal grouping by usual methods of preparation may have been 

 referred to that species. Certainly the list of forms grouped under 

 that name by authors is long, and perhaps loose. N. viridis, however, 

 which frequently accompanies this species in typical condition, always 

 shows in marked contrast, as regards shape of valve, less parallel 

 costae and more curved, complex raphe, as well as in the absence of 

 grouping into motile chains of four. N. viridis var. commutata Grunow 

 has dissimilar valves; or, if we shall follow Cleve 2 in his rather puzzling 

 rearrangement of this group of forms, var. commutata differs in its 

 more approximate costse (10 to 12 in 10 ;x) and their divergent tendency 

 in the middle portion. Still following Cleve, var. fallax, Cleve also 



2 P. T. Cleve, Synopsis of the Naviculoid Diatoms, II, p. 91. 



