- APRIL 25, 1884.] 
nearly four feet at the top, becoming narrower towards 
the base. Mr. Thompson has also observed the same 
phenomenon on sharp rocks and sticks. Ice-banners 
are evidently formed from the vapor of passing 
SCIENCE. 
523 
clouds; and an observer favorably situated might 
watch their formation and growth. He thinks that 
possibly the base of a cloud-banner might be found 
to be an ice-banner. 
RECENT PROCEEDINGS OF Ss CIENTIFIC SOCIETIES. 
American society of civil engineers. 
April 16.— A paper was read by Hamilton Smith, 
jun., upon the temperature of water at various depths 
in lakes and oceans. The results of observations 
upon bodies of water in California, in the eastern 
States, and in Switzerland, were collated, and also- 
the temperatures obtained in deep-sea soundings; all 
of which show that very slight variations in tempera- 
ture occur at great depths, and also that great varia- 
tions in surface-temperature affect the deeper waters 
only after a long interval, and that even in compara- 
tively shallow reservoirs there is great uniformity in 
temperature, at even moderate depths, as compared 
with the variations in its surface. 
Brookville society of natural history, Indiana, 
April 8.— A. W. Butler presented a paper upon 
some explorations among the ruins of San Juan Teo- 
tihuacan, near the City of Mexico, illustrated by 
maps of that region, showing its topography. He 
described the appearance of the pyramids, the ‘ House 
of the sun,’ and the ‘ House of the moon,’ and gave 
the results of his investigations of the manner of their 
construction, and the excavations near them. A de- 
scription of the so-called ‘Micoatl,’ ‘Path of the 
dead,’ and its relation to the ‘ House of the moon,’ 
were given. In conclusion, he mentioned a large sac- 
rificial stone found near the ‘House of the moon,’ 
which he illustrated by drawings of the front and 
top. —— Edward Hughes read a short paper upon the 
rats of Franklin county. ——A. W. Butler gave a 
short paper on the tornado of March 25, which he il- 
lustrated by maps, showing its course through the 
eastern part of Franklin county (Indiana), and the 
destruction it caused at Scipio, Ind. 
Torrey botanical club, New York, 
April 8. — Mr. Arthur Hollick read a paper upon 
autumn forms of the genus Viola. While engaged 
in studying the cleistogamous flowers of V. cucullata 
and V. sagittata, many other species were brought 
under notice, and important differences remarked in 
leaf, flower, and stem, which do not seem to have been 
previously reported. V. cucullata and V. sagittata 
‘are connected by every conceivable intermediate form 
of leaf variation and superficial characteristics, and 
VY. palmata also connects with the former by insen- 
sible gradations. V. cucullata and its varieties are, 
however, distinguishable from either of the others 
mentioned by the decumbent habit of the cleistog- 
amous flowers. In V. sagittata, on the other hand, 
the cleistogamous flowers are invariably erect. For 
‘Some time it was difficult to know whether V. palmata 
was allied to V. cucullata or VY. sagittata, but the 
appearance of the intermediate forms points to the 
former as the type. The three species of white 
violets —viz., V. blanda, V. primulaefolia, and 
V. lanceolata — are very closely allied, intermediate 
forms between the latter two being impossible to place 
accurately with either species. All three produce 
runners or stolons late in the season; but in V. blanda 
these runners are merely roots, being almost entirely 
under the surface of the ground, slender, and produ- 
cing few or no leaves, and no cleistogamous flowers. 
The flowers grow from, or close to, the main root- 
stock, and are more or less decumbent. YV. primu- 
laefolia has the longest runners, some as much as 
twelve inches in length. They are comparatively 
stout, run along the surface of the ground, and are 
mostly leaf and flower bearing throughout. V. lan- 
ceolata will probably have to be referred to the same 
species as the latter. An important point to be noted 
is, that V. primulaefolia and V. lanceolata almost in- 
variably grow in company with each other, while V. 
blanda generally occurs alone, and in different loca- 
tions from the other two. These violets have three 
methods of propagation, — by petalous flowers in early 
spring, by apetalous flowers in the autumn, and by 
runners rooting at the nodes or joints. V. odorata 
produces both leafy runners and cleistogamous flow- 
ers; but the flowers are clustered around the main 
stem, instead of being on the runners. They are 
depressed upon short peduncles, and are sometimes 
almost subterranean. In VY. canina, var. sylvestris, 
the cleistogamus flowers have peduncles not more 
than two inches long, generally less, while the others 
are from three to four inches in length. Also, while 
in the spring flowers only one starts from each axil, in 
the autumn forms there are usually two or more. 
V. pedata apparently never produces cleistogamous 
flowers, but very frequently blossoms a second time 
in the autumn. Specimens were collected as late as 
Noy. 5 in full bloom. —— A committee was appointed 
to prepare resolutions urging the necessity of legis- 
lative action in regard to the preservation of the 
Adirondack forests. 
Colorado scientific society, Denver. 
April 7. — The committee on artesian wells in the 
neighborhood of Denver made a preliminary report, 
outlining the basin within which the known flows 
might be obtained, and giving calculations as to 
the amount of water available. Mr. E. LeNeve 
Foster described a possibly new mineral from Mexico, 
having approximately the formula, 4 Ag.S.6PbS. 
5 Bi,S3. It occurs as a massive cement to a granu- 
