Molhr's Diatomaceen-Prohe-Platten. Bij E. W. Morley. 225 



in the same place as was permitted by the fact that the number 

 of striae counted was usually varied. Here the variations in the 

 numbers given for the same diatom frustule are mostly due to 

 errors of observation. These errors were kept as small as seemed 

 necessary for the purpose ; that they could be made much smaller 

 if required was proved by the result of extra care taken with diatoms 

 Nos. 19 and 20 on Platte No. 258, where the variations from the 

 mean were all very small. More time was taken for these two 

 than for any others of the whole two hundred ; an hour and a half 

 of lamplight and an hour of sunlight were commonly sufficient for 

 measuring the diatoms of one Platte. 



In the table following, the first column gives the name of the 

 diatom with its number ; the second describes briefly the part of 

 the diatom where the measurement was made ; at the end of the 

 table certain notes specify this part more accurately in certain cases. 

 The other columns give the measurements for the Platten named at 

 the head of each column; the first number gives the number of 

 striae counted in that measurement, and the second gives the number 

 of striae in the thousandth of an English inch. Under the names of 

 the diatoms in this table are given the extremes for each one of the 

 twenty in the series. In this statement of the extremes, it is 

 assumed that the measurements for Nos. 1 and 2 are trustworthy 

 to a unit in the first decimal place, those for the diatoms from 

 No. 3 to No. 9 inclusive are considered trustworthy to half a unit ; 

 for the remaining diatoms the whole numbers correspondmg most 

 nearly to the fractional results of measm'ement are given with no 

 expression of opinion as to the limit within which they are trust- 

 worthy. It may be stated that the widely divergent result for 

 Cymatopleura elliptica on Platte A is not due to error; the 

 result was re-examined after it was noticed that it was thus 

 divergent. 



Of course the amount of variation shown by these measurements 

 is no index of the extreme variation in diatoms of any one of these 

 species from different sources. But the variations found may be 

 fairly assumed to show the order of magnitude of the variation to 

 be expected among diatoms selected with care from the same source. 

 It may be expected, for example, that samples of Amphipleura 

 pellucida, carefully selected from the same stock, may vary as the 

 numbers 52 and 95; while it could not fairly be expected that 

 under these conditions there should be any such variation as from 

 80 to 110 or 120. The results to be obtained by the use of those 

 diatoms on the Platte which are shown by the table to vary least 

 when selected as Moller selects them, may be compared with no very- 

 wide limits of error, as far as such results depend on the fineness 

 of striation of the diatoms in question. 



R 2 



