No. 640] VARIATION AXD HEREDITY IX /.r/V.Yr.s' 435 



posed places and usually in poorer soil. No cultures of 

 different-sized j)lants of the same si)ecies have yet been 

 grown and critically compared. Some casual observa- 

 tions, however, indicate that there are prohahlx lici-itable 

 size differences in apricus. 



Keference to the s])ecific descriptions will show that 

 some taxonomists have been inclined to base classifica- 

 tion in part on the coh,)r, shape and markinji's of the seeds. 

 Our observations would indicate that this is not a very 

 safe criterion, particularly in reference to color and 

 markings. Seeds of i/aiiiis arc tlie largest. They could 

 usually be picked out of a mixture with apricus 1»y size 

 and shape, but not by tlie color oi' inarkiiiu-<. Xo clinr- 

 acter is more su})ject to variation than this, 'riiis does 

 not mean that the charactei- is a thn-i na i iim' on;', hiit that 

 there are a great many differently colored and marked 

 seed varieties. All the seeds of any one ])lant are alike 

 (with certain exceptions to be noted in a later para- 

 graph), as one would anticipate since tlie color and mark- 

 ings are seed-coat characters and hence genetically all 

 alike on any ])ai-tieular ])lant. 



Dark bine lunni. nn.l nfun»s p'ant^ U■A^r .laik ^-eds, 

 whites and pinks lia\-e liaht seed^, liubl hlues and -tri])ed 

 whites have seeds of mtei inedlate coloi-. Pinks and lidit 

 blues have a yellowish-hrown tone and are flattened and 

 without other mark iim-. Stiiped whites ai-e lonixer and 

 thicker in proportion to the width, of either a bluish or 

 yellowish tone with a conspicuous crooked line on cnther 

 side forming a continuous ring about the shorter perim- 

 eter of the seed. This ring does not occur on the seeds 

 of pinks or whites. Tt is ]H-esent on some races of dark 

 blues and absent on otlMM->. The -tatenient< \\\<\ made 

 apply particulai'lv to tlioH' races wliicli have been culti- 

 vated for some time in the cxprrlnK'tit -arden. Field 

 observation coniii'ins them in pari, but rc\c;d.s a cnn^id- 

 erable number of otli<'r \ariations in coloi- and marking 

 which have not yet been cn!li\ated and about the iz-enetic 

 behavior of which nothinii- is yet known. The only case 



