368 JOHS. H(UE PETKRSKN 



of tilis 1 mixed wilh ;i (Irop ol' walci- on :i cover glass, diied il, 

 and llicii heated il lo red heal oii a plaliiumi plale. Oii llie eover 

 glass I i)laced a droi) of purilied styrax dissolved in henzole and 

 alcohol, the cover i>lass Nvas slii^hlly healed unlil Ihe solvcnls had 

 evaporaled, and Ihen placed face downwards, on a slide, in Ihis 

 way I ohlained permanent preparalions, in which the structure of 

 the Irustulcs could be closely examined, and where the })lace of the 

 individnal fruslules was praclically unchangeahle. Ilowever, the object 

 ol my investigalions was not only to ascertain what Diatom fruslules 

 xvere to be found in Ihe earth, bul in Ihe Ihst place to lind out 

 what species lived on Ihe ground. Hence, alter determining the 

 species in Ihe styrax preparalion, I had lo try and iind out from 

 a preparalion in waler of the same material what species had ac- 

 tually lived on the spol, and il", possibly, some of the fruslules found 

 mighl belong lo species thai had nol lived there at all bul had 

 merely happened to be carried to Ihe place where the sample was 

 taken. Thai this mighl occur we must be preparcd to expect con- 

 sidering the great resislance of the Dialom fruslules. Il will be seen, 

 under the individnal species, that the samples in which the species 

 occur are sometimes marked wilh an L in front of Ihe number. 

 This is to denote thai upon examination of the waler prei)aralion 

 one or more specimens of this species conlaining »endochrome« 

 were found in the sample. Since the liner structure of the valve is 

 not visible in a waler preparalion, there is a possibility of making 

 mistakes. However, the probabilily of such seems lo me greatly 

 diminished by the faet that I always beforehand examined a styrax 

 |)re])aralion of the same sample and so knew what species occurred 

 in the sample. Moreover, il will not, probably, be possible in such an 

 examination to see all the species that have lived in Ihe place wilh 

 their cell conlenls, for numerous empty fruslules will always be found 

 in a Diatom vegetation even if il is (juite luxurianl. This will perhaps 

 be the case especially in Dialom growlhs on Ihe ground where 

 many specimens will doublless die in unfavourable (i. e. dry) periods. 

 As among the Phanerogams, so also among the Dialoms Ihe 

 species seem very heterogeneous. In some parts of Ihe system, 

 very small and narrowly delimited species occur, while in olher 

 parts there are jnore extensive sj)ecies wilh numerous varielies and 

 forms. This dissimilarily is no doubt in the main due to our fading 

 power to estimale the greater or lesser signilicance of the systematic 

 characters employed, hence we are unable to form any decided 



