1H91.] MICROSCOriOAL JOUltNAL. 99 



these are the several characteristics : Plcurosignia is curved like an S ; 

 Staii7-oncis lias a cross-lxir on both valves, Achnanthcs only on one. 

 In Phinularia the cross-lines resemble the teeth of a comb ; Amphora 

 is inflated, and the frustule usually shows two knobs in the middle of 

 the flat side ; Gomp/ionctna is wedge-shaped ; Cocconeis is elliptical, 

 and very broad for its length ; Cocconcnia is bent ; Naviaila has acute 

 ends, no cross-bar, and the lines are resolvable into dots ; Cocconetna = 

 Cyvibclla with a stalk, and AchtiantJiidinm == Ac/itiatithcs without a 

 stalk. 



But suppose your diatom has no nodule. If so, it belongs to the 

 other division, which includes the following i8 genera: I^pit/iemla, 

 Kjinotia. irhnantidiutn^ Aleridion^ Diatoina^ Asterionclla^ Fragi- 

 laria^ Dcnticula., Odo?itidium, Cyclotella^ Mclosira, Stcphanodisctis^ 

 Catti pylodi scus ^ Surirclla^ Cymatopleura^ Nitzschia^ A7iiphipleiira., 

 Syncdra. Of these Meridion can be readily told by the spiral form of 

 its united frustules. Now, as to the rest : 



Is your diatom shaped like a bozvP It is Hpithcniia^ Eunotia^ of 

 Uimantidittm. If it is ribbed it is the first ; if it has knobs at the ends, 

 the second ; if the frustules form a filament, the third. 



Is it rectangular in front viexv^ and are several of the frustules con- 

 nected? It '\^ c\\\-\c\ Fragilaria, Odontidinin^ Diato7na^ ox Aster io- 

 nella. Fragilaria forms a straight filament ; Diatonia a zigzag chain ; 

 Aster ionclla a star; in Odontidium the filament has a slightly uneven 

 outline ; in Dcnticnia usually only two frustules are united. 



Is your diatom disk-shaped? It is either Cyclotella^ Alclosira, or 

 •Stcphafiodiscus. And if it forms a filament of jointed cylinders, it is 

 the second, while a fringe of little teeth will make it the third. 



Are the inargi?ts odd? It is either Campylodisctis^ Surirella^ or 

 Cymatopleiira. If the frustule is twisted, it is the first; if the valve 

 has a longitudinal line, it is the second : dark bands on the valve dis- 

 tinguish the third. 



Finally, is it very long hi proportion to its width? It is either 

 Nitzschia^ Synedra^ or AtnpJiiplcnra. If the longitudinal line is not 

 in the middle, it is the first ; if the sides are nearly parallel, the second ; 

 if the longitudinal line has terminal expansions, the sides of which are 

 ))arallel, it is the third. 



Or the beginner may start equally well with the fossil forms, if he 

 has suitable slides for study. In that case he will have but 25 genera, 

 and even 5 of these may be passed over as being classed as doubtful or 

 obscure by the Micrographic, namely : Amphicampa^ A?nphipentas^ 

 Mastogonia^ Svniholophora^ and Stephanogonia. Divide the rest 

 thus: Breast-shaped — Ilercotheca^ GotiiotJiccium : Heavy lines or 

 seams — Bihldrium^ Stylobibliiim : Disk-shaped — Actinocyclus, Ac- 

 tinoptychtts^ lleliopelta,, Coscinodiscus^ Arachnoidiscus^ Astero- 

 IcCfnpra^ Asteromphalos^ Halionyx^ Odontodiscns^ Omphalopelta^ 

 Systephania : with appendages — Hupodisciis^ Auliscus, //emianlus^ 

 Periptera^ Dicladia. 



Certainly, thus divided, this is no longer a formidable list, and you 

 notice that you are dealing with forms quite distinct from those of fresh 

 water. Those were mainly long forms ; these are chiefly rotmd. 



In the first division marginal hairs distinguish Ilercotheca. 



In the second, Stylobihlium is round, antl the lines form zones, in 



