PLANTS OF WESTERN LAKE ERIE. 
75 
are of this nature, and Myr iop hyll/um also passes the winter in this way. Winter 
buds were common on three species of narrow-leaved Potmnogetons , P. zoster wfolius, 
P. pusillus , and P. freisii (pi. 15, tig. 2). Potamogeton lonchites sometimes propa- 
gates by means of short branches, which produce buds at their ends. Roots and 
leaves grow out from these buds, and the result is a small plant, ready to root and 
grow whenever it is detached from the parent plant. 
ANALYSIS OP SOIL SAMPLES. 
The samples of soil collected at Put-in Bay, East Harbor, and Sandusky Bay 
were analyzed by the Division of Soils, U. S. Department of Agriculture. The 
results of the analyses are given in Table I. The number of samples is not sufficient 
to make general deductions possible, and therefore I shall merely indicate the 
direction in which the results seem to point, leaving it to future work to establish 
the relation, if any exists, between the texture of the soil and the plants growing 
upon it. 
By reference to Table I, it will be seen that, as a rule, the soils on which plants 
occurred in abundance were composed largely of fine sand and very line sand, and 
contained relatively little silt, fine silt, and clay, while the soils on which few or no 
plants occurred, although the depth of water and other physical conditions were 
favorable, were composed largely of silt, tine silt, and clay, and were poor in fine 
sand and very tine sand. The other items are of no practical importance, the amounts 
of gravel, coarse sand, and medium sand being very small, while the amount of 
organic matter is not at all regular, being relatively large in all samples from places 
where no plants grew and irregular in the other samples. Of all the samples taken, 
six must be excluded from the comparisons on account of other factors coming 
prominently into play. Sample 1, from Gibraltar Bar, is not comparable with the 
others, both on account of its mixed character and the exposed position of the bar, 
and the three samples from the open lake, Nos. 3, 4, and 7, were taken at a depth of 
33 to 36 feet, and hence can not be compared with samples taken at depths not 
exceeding 7 feet. The two samples, 11 and 12, collected on the lake shore,' were taken 
to determine the cause of the presence or absence of Scirpus. All the other samples, 
ten in number, were taken from places where the depth of water ranged from 6 inches 
to 7 feet, and where all the other physical conditions were nearly similar. 1 have, 
therefore, divided these ten samples into two sets, six in one and four in the other, 
and have added together for each sample the percentages of fine sand and very tine 
sand to make the first column of Table II, and the percentages of silt, fine silt, and 
clay to make the second column. The six samples of the first set represent localities 
well stocked with plants, while the four samples of the second set were taken from 
bottoms either bare or on which but a few plants were growing. 
The agricultural value of soils is largely determined by their power to retain 
water — sandy, dry soils being good for early truck crops, but almost useless for the 
heavier late crops, as wheat or corn; while soils containing much clay retain water 
better and are consequently later and colder but more valuable for wheat and grass 
crops. As shown in pi. 2, Bulletin No. 5, Division of Agricultural Soils, a typical 
truck land contains 79.69 per cent of medium, fine, and very fine sand and 14.36 per 
cent of silt, fine silt, and clay. It is somewhat similar, therefore, to the sample of 
