224 BOTANICAL GAZETTE [MARCH 
relation to temperature, and with those given by BAtts for rate 
of growth of sore-shin fungus in relation to temperature. Omissions 
in the tables of coefficients are due to lack of comparative data at 
certain temperatures. 
Figs. 1, 2, and 3, included as representative graphs, show the dif- 
ference in total percentage germination of seeds collected at the 
same time in one locality, and germinated at the same time at the 
same temperature but with seed coat intact in one lot, abraded by 
grinding with sand in another, and carbonized with H,SO, in the 
third lot. This difference in total germination is greatest at low 
temperatures. Nevertheless, the coefficients for rate of germina- 
tion computed from these curves in all lots run remarkably similar 
(tables I-III). 
Fig. 4 shows the striking restricting effect of the coat at high 
temperatures, an effect paralleling the magnitude of the restricting 
effect of this coat at temperatures near the minimum temperature 
for germination. The removal of coat effects, either by grinding 
the seeds with sand or by treating with H,SO,, allows much more 
rapid germination at the same high temperature (42° C.) for the 
Washington seeds. Effects somewhat paralleling these were ob- 
tained for Indiana seeds at 46.1° C. In fig. 4 the effect of grinding 
the seeds with sand nine weeks is shown to be greater than the effect 
of grinding for six weeks. Concomitant with this slowness in 
germinating is the lack of anthocyanin in the seedlings developed 
at these high temperatures, and again apparently lessened amount 
in the seedlings developed at the lowest temperatures at which 
germination took place at all. 
Conclusions 
1. The coefficients relating rate of germination of seeds of 
Amaranthus retroflexus to temperature grade from high values as 
10.01 (table II) at low temperatures, to low values as 0.001 (table 
II) at high temperatures, thus paralleling the coefficients relating 
rate of growth of seedlings and sore-shin fungus to temperature 
worked out by Lerrcu, LEHENBAUER, and BALLs. | 
2. The general trend of these coefficients is the same for seeds 
only partially after-ripened, and for those with coat effects almost 
