230 ANNUAL REPORT OF Off. Doc. 



tober, 1898, nearly 68 per cent, were dead the following spring. Death 

 seems to have been caused through the roots rather than by the 

 tops, for many of the killed trees stood until September with green 

 tissue just above the surface of the ground and in a few cases running 

 half way up the trunks, while their roots were decayed to a powder. 

 It should be noticed here too, however, that in the thinning opera- 

 tion referred to, the superfluous trees were cut out by severing the 

 tap root about two inches below ground, and in very many cases these 

 roots left in the ground sent up sprouts the present summer, strong 

 and vigorous, even two or three sprouts from one root. It is a sur- 

 prise that these roots were not killed. 



Table II. The seeds of the bur oaks (all from native trees) showed 

 great variation in size and shape. Those from Vermont were the 

 smallest, being one-half inch in diameter and three-fourths inch long. 

 Those from Oklahoma were the largest, measuring as much as one and 

 one-half inches in diameter and two inches long. The same quantity 

 of acorns was sown in each case, but the number of plants obtained 

 was very small, except in the case of Vermont. It is a fact that the 

 larger the acorns, the smaller the number of resulting plants, though 

 this may not always be the case. The number of plants in this series 

 of cases is too small to warrant definite conclusions. 



Table III. The seeds of the eleven samples of hackberry were quite 

 uniform in size, that is J- inch in diameter with the exception of those 

 from Texas and South Carolina, which were small enough to be 

 the seeds of the shrubby variety of the south (Celtis occidentalis 

 var puinila). The foliage on the South Carolina plants is much 

 smaller than that of the others, though the young trees are fully as 

 strong. The earliest plants appeared through the ground May 5 

 and the last on July 9. The seed from North Carolina failed entirely. 

 All the plants from the Texas seed, fourteen in number, were killed 

 in their first winter, arid those from Oklahoma, originally 117, being 

 among the most vigorous hackberry trees were thinned to five good 

 specimens by October, 1898, and all were killed by the severe winter 

 following. The seeds of all the hackberries were from native trees. 



Table IV. The honej ..ocust seed was all alike and apparently fresh 

 and good. They were jtxom native trees with one exception, that from 

 Pennsylvania. One-ht<f of the samples were planted in the fall, 

 as soon as received, but the remainder reaching us in midwinter, 

 were stratified and planted in the plot April 14, without soaking in 

 water, as is often recommended for such seeds. A good percentage 

 of germinations occurred promptly though many others made their 

 appearance throughout the first summer. All the trees have main- 

 tained a uniform growth and development, showing the least varia- 

 tion of all the species of trees grown. It is not necessary to state 

 here observations which are fully indicated in the tables of records. 



Table V. Out of thirteen samples of box elder seed, three were 



