VARIATION IN TRILLIUM GRANDIFLORUM. 171 



may be increased by natural division of the rootstocks of the 

 two-stemmed individuals, and perhaps also by cross pollination, 

 as in many of the very abnormal forms one or more of the 

 stamens produce pollen, which is probably potent. In fact 

 among the plants examined, in only rive flowers was it noted 

 that none of the stamens were pollen bearing. 



The plants here described were collected near Syracuse, N. Y., 

 in a wood of second-growth timber. The soil, which overlies 

 a limestone formation and which is more or less intermixed 

 with limestone rocks, is a rich leaf-mould on top and a compact 

 clay loam beneath. The rootstocks usually rest on the clay 

 and most of the roots penetrate into it. The richness of the 

 locality in trillium individuals is only poorly shown, by the first 

 illustration. In a strip of territory hardly a quarter of a mile 

 wide and less than a mile long normal plants occur by the 

 hundreds of thousands and abnormal ones by thousands. At 

 some spots barely half a dozen abnormal forms can be found 

 among a thousand plants, while at a nearby spot from ten to 

 fifteen out of every hundred will show coloration of the petals 

 with the accompanying variations of the other parts. On the 

 whole, probably at least one per cent of the plants shows 

 abnormal variation. 



While the measurements given indicate approximately the 

 size of each part, they do not of course indicate the shape of 

 the outline. This varies to some extent in the cases of the leaf 

 blades and sepals, but very conspicuously so in the case of the 

 petals. Thus, as the photographs and table of measurements 

 show, plants 13 and 143 have petals more than three times as 

 long as they are wide, while numbers 22 and 31 are nearly as 

 broad as they are long. Numbers 84, 105 and 163a are just as 

 broad as long, while 163& is broader than long. But, however 

 much the outline may vary, the petal never loses its pointed tip. 

 In some of the specimens examined it was in a deeper notch than 

 shown in plants ill and 22. It comes more nearly being 

 obliterated in extremely broad-petaled plants of the normal sort, 

 such as number 12, than its does in any of the greatly abnormal 



varieties. 



In the following table all the measurements are in millimeters, 

 the greatest width of the organ being given first and then the 

 length. When two figures are given in the column "Length 

 of ovary," the first refers to the length of the stalk or stem upon 



