44 Rhodora [Marcu 
I should say that the balls were composed chiefly of the debris of the 
chestnut burrs, with some admixture of the veins of rotted leaves. I 
have found them from an inch in diameter, in which case they are 
flattened and loosely compacted, up to six inches, when they are per- 
fectly spherical and closely compacted. The northern shore of the 
pond is shallow for a long distance out, and the prevailing southerly 
winds of summer would roll the waves on the shore. The summer 
of 1898 was one of extreme low water; the following years the water 
rose several feet, and I was unable to find a single Burr ball. 
Mr. Riddle is evidently not certain as to the composition of the 
Sandy Pond balls, and while it is entirely possible that some com- 
posed of the debris of chestnut burrs may occur in the pond, it is 
likely they are all of Eriocaulon, as the other observers have stated. 
Certainly the several specimens 1 have received from Mr. Hosmer 
and Mr. Richardson, and one of which is figured on the accompany- 
ing plate, are composed of Eriocaulon almost exclusively. 
So much for the information 1 have obtained about the balls from 
Sandy Pond. Meanwhile 1 was shown at Fredericton by Professor 
L. W. Bailey, of the University of New Brunswick, a very fine speci- 
men, of similar formation but very different material, said to have 
been found in Kedron Lake in the southwestern part of York County, 
New Brunswick. Following up the matter 1 found that it had been 
obtained by Mr. Wellington Davis of Brockway, N. B., from Little 
Kedron Lake, and on application to Mr. Davis he sent mea fine 
specimen together with the following account of its formation : — 
It is found in the north end of Little Kedron Lake in a small cove. 
No wind can strike the cove but from the southeast. It is surrounded 
with fir and spruce which hang over the water. The bottom is a 
clear sand. The spills drop from the fir and spruce and lie at the 
bottom. Then the water washing from side to side forms the ball. 
No heavy swell comes in there. These balls can be found in no 
other place in the Little Kedron Lake nor in Big Kedron Lake. 
Sometimes we have found them from six to eight inches through. 
This account, with other matter relating to this subject (practically 
all included in the present paper), together with a plate showing the 
ball beside one from Sandy Pond, were published in the Zducational 
Review (St. John, N. B.) for August, 1904 (page 51), and it appears 
in somewhat altered form in the Bulletin of the Natural History Soct- 
ety of New Brunswick, No. XXIII, 1905. As Mr. Davis states, the 
balls consist almost exclusively of the spills (v/z. leaves or needles) 
of the spruce, which are interlocked in a very compact and fairly 
tenacious mass. It is very remarkable that such smooth objects 
