M 



JOURNAL OF CONCHOLOGY, VOL. l6, NO. I, JANUARY, I919. 





Date. 



Number. 



Mean. 



Standard deviation. 



a 



Aug., 1913 



204 



11-356 



.. 0-741 



(3 .. 



3V7/14 



209 



I I '490 



0-804 



7 •' 



23/8/14 



.. 296 .. 



11-407 



.. 0-707 



S .. 



29/8/14 



•• 75 •• 



II393 



0-672 



e 



30/8/14 



.. 132 .. 



11-450 



0-760 



Standard deviation. 



0-642 

 . 0-652 



0-700 



The largest difference is a/^ = 0-134 m.m., which is less than twice 

 its standard error, the others are still less significant ; evidently there- 

 fore, there is no demonstrable difference between the five samples. 



From another locus (H) there are three collections — 



Date. Number. Mean length. 



a ... 2/8/14 ... 100 ... ii'478 



f3 ... 23/8/14 ... 306 ... 11-485 



y ... 28/8/14 ••• 57 ••• ii'4oo 

 The largest difference is f3/y = 0-085, 0-8 times its standard error. 

 From a third locus (B) there are two lots — 



Date. Number. Mean length. Standard deviation. 



a ... 23/8/14 ... 66 ... 11-223 ••• 0726 



13 ... 28/8/14 ••• 86 ... 10-991 ... o-6o2 

 The difference is 0-232 m.m., its standard error o-iio. 

 From a fourth locus (C) there are also two lots — 



Date. Number. Mean length. Standard deviation. 



a ... 9/8/14 ... 79 ... 10-162 ... 0-692 



/3 ... 27/8/14 ... 44 ... 10-382 ... 0-747 



The difference here is only 1-6 times the standard error. 



It occasionally happens that samples derived, as far as may be, 



from the same locus, show significant differences. Thus one of the 



collections (y) from locus A was actually made in eight small lots, 



which gave the following results when treated as separate samples — 



I 

 2 



3 



4 



5 

 6 



7 

 8 



The largest difference is 2/5 = 0-477 m.m., which is 3-7 times its 

 standard error; the next largest is 2/4 = 0-455 m.m., which is 3-0 

 times its standard error; the next biggest difference 1/5 =^0-406 

 m.m., which is 2-9 times its standard error; the other differences are 

 all much less than three times. Out of 28 possible comparisons 



Number. 



Mean. 



Standard deviation. 



31 



I I "200 



0-578 



51 



II-I29 



0-629 



27 



11-504 



0-791 



31 



11-584 



0-689 



47 



11-606 



0-637 



31 



11-494 



0-693 



41 



11-317 



0-634 



37 



11-519 



0-651 



