r r 



■1 



MANX— DIATOMS OF TITE ALBATROSS VOYAGES, 241 



C, Ag. I agree AviUi Caytracane that this union i^ to be rej(^cleJ< For, as he points 

 oul/^ ihe present y^ejuis is never met with growing as Melosira invariably ck)es — in long, 

 solid, thread-like hlanients. And although a f(^w specie's of Melosini ^jresent a rougli- 

 ened center of the valve slightly resembling the '' uinliilieiis-'' ot Podosira, it is not 

 eharaeteristie of the geTius, and wliere it does occur tlie blotch is insignificant, never 

 shows a sutural line of separation from the rest of the valve — in short, is an altogether 

 different thing. So that although individual valves of Melosira and Podosira maybe 

 found somewhat to reseml)le each other, the tyt>ieul forms ajid tlu^ modes of growth 

 are so widely different that their union would ereat<^ niurh confusion and adtl unneces- 

 sarily to tin; already unwiiddy bulk of iUdosira. I think the only author following 

 tills 2>lan (^f Sini ill's is Lagerstedt. 



Kiit/ing wislunl to uK'rgf^ Podosira in Oy('lf>^'' Ha; but its generally hirg(^r size, more 

 delicate structure^ convex valves, and, abovi^ all, its prominent c^Mitral umbilicus, 

 make such a classillcatittn imj)ossible. 



Ralfs ^ distinguishes between Podijsira and Ilyalodiscus by stating that in the latter 

 the valv(\s are flat. This distinction is not true, and would not be impoilant if it were. 

 Xone of the species or varieties of Ilyalodiscus are flat, though in some cases the con- 

 vexity of the valv<* is slight. IJoth have the characteristic central luubilicus, bounded 

 by an evident suturt*. liolh shcnv the same sort of markings, that of tlie umbilicus 

 being l>lotched or rugos<% that of tin* rest of the valve being delicate beading, gt^nerally 

 arnjnged in curved liiu*s. producing tin? appearance of ''watch-cas(* milling." Both 

 grow in the same way, one or a few frustules joined by slutrt, stout, gelatinous stipes. 

 Some spt^cies prc^vitmsly classified as Podosira ((>r ^^till so classified by anyone who runs 



the genus into Melosira) are wholly (U'Stitute of 1 he before-mentioned umbilicus; but 



that it lias been generally lookinl upon as an esscaitial mark of Podosira Is not only 

 expressly stattnl l>y ilalfs, but i'vident from hundreds of figures of both Po<l(»sira and 

 Ilyalodiscus. <'h*ve & (irunow^'get into the difhculty common with nuiny authors of 

 trying to <listinguish l>etween the tw(» gen(*ra on the basis of ihe um|)ilicus. vVfler 

 having stated that '* Iltjalodiscits hat ein mehr oder wenigcr scharf abgesondertes Cen- 

 trum," it is adrnilted that this is not univ<'rsallv so. as in varielies of //. ,^f(btiUs Pail., 

 '\so dass sie sidi in di(^^er llinsicht gar nicht von numclu'ii l-'ormen der Podosira 

 viaxwui untor.srh(Md(^n !"' In anotluT i>hice, aft(^r staling thai l*odt)sira nuiy lie looked 

 uptm as without an umbilicus, the writers achl thai in the case of /'. intuitita and l\ 

 ambigua '^ein ganz entschiedmer Pml)ilicus vorkommL*' 1'lie whoh- discussion is a 

 good illustration of the imi^ossibility of liolding (hese two g(*n4'ra sej)arate. It is c-er- 

 tainly well known to all who ha\e examined gatlu^rings rich in Podosira iliul in llut 

 same species and from ihe same locality the umbilicus varies gi'eatly in size and dis- 

 tinctness, and individuals are not hard to find where scan civ a I race of it remains. 

 Thus in llu^ five species of Pod<tsira given in ihe li, L. Smith ty|)e, four are wilh v<'ry 

 strong umbilici; one^ P. vionlagnfi Kiitz., a possil)le ^b'losira, withoul. In P. hor- 

 moidcs (Mont.) Kiilz. ihe uml)ilicus is generally strong^ in some cases quite small, and 

 in at least a dozen valves on this single strewn slide almost impossible* to see. Van 

 Ileurek '^ says: "The Ilyalodiscus ar(^ not essentially different from Podosira, exc(^pt 

 in the uml)ilicus, which is more or less distinct aecordiiif^^ to species." 



If, theref(jre, we remove from this genus the forms that evidently behjng to other 

 genera, chiefly Melosira and Cosciuodiscus, wt^ niay defuie it as follows; 



Frustules growing singly or in a series of a few membt'rs, atta<^"hed to a suj>port and 

 to the next in Ihe series l>y sliort, stout, gelatinous stipes, cenl rally ])lac(Hl, these 

 usually causing a pronoun<*(sl ''umbilicus-" or rugose scar situatiul in the center of 



^^Castr. Rep. Yoy. Chall. i5ot. 2: 131). 18S(;. 

 ?>Pritch. Hist. Infus. ed. -1. 814-815. JSSI. 

 ^Sv. Vet. Akad. llandl. 17-: 115 IPi. ISSO 

 d Van Ilcur. Treat. Diat. 448. 180(i. 



