34 BOTANICAL GAZETTE [JANUARY 
treatment may have on sap-flow. For comparison with my 
observations on Cyperus, ROSHARDT’S experiments on the grasses 
are of particular interest. Experiments were made upon Bromus 
sterilis, Bromus hordeaceus, Dactylis glomerata, Poa pratensis, 
Agropyrum repens, Glyceria plicata, Arrhenatherum elatius, Lolium 
perenne, Cynosurus cristatus, Secale cereale, Miscanthus poly- 
dactylos, and Bambusa aurea. With the exception of Secale, 
Miscanthus, and Bambusa, 5-10 similar halms were killed for the 
same stretch. Although the grasses agree generally in the mor- 
phological structure, he finds that their response to the treatment 
was widely different. He was unable to observe distinct wilting 
of the leaves except in Bromus hordeaceus, and then only in direct 
sunlight. In other species of the grasses, yellow-red spotting to- 
gether with a yellow discoloration of the leaves appeared, which was 
followed by shriveling. The leaves began to dry first at the tips and 
fading began from below upward, the upper leaves drying last. The 
flowers and fruits remained fresh longer than the leaves, and the 
halms dried last. The leaves of Miscanthus wilted from above 
downward on the stem. The lack of water was often indicated 
by a rolling of the leaves, which occurred in Agropyrum and 
Bambusa. He tabulates the results in cases in which a length 
of ro cm. of the halms was killed in order of their drying: Bromus 
sterilis (dry in 17 days), B. hordeaceus (16 days), Bambusa (13 days), 
Dactylis (12-14 days), Glyceria (10 days), Poa (9 days), Lolium 
(9 days), Arrhenatherum (about 8 days), Miscanthus (7 days); 
and Secale (about 7 days). When 20 cm. of the halms are killed, 
the following order results: Agropyrum (12-13 days), Bromus 
hordeaceus (12 days), Bambusa (10 days), Glyceria (8 days), Poa 
(5-7 days), and Miscanthus (2-5 days when 22 cm. were killed). 
It will be seen that the leaves of Bromus hordeaceus remain fresh 16 
days when to cm. are killed, while its leaves dry more quickly 
(12 days) when 20 cm. are killed. He feels justified in concluding 
from these results that the longer the killed stretch the sooner 
the leaves above indicate a deficiency in the water supply by drying. | 
RosHArp1’s tables show that killing the same stretch of the stem — 
of nearly related forms of the same genus produces very different 
effects in the time of wilting of the leaves. He has compared 
