110 MICHIGAN ACADEMY OF SCIENCE. 



as well as ditfereut individuals differed considerably. In most cases there 

 seems to be no regularity in the variation but the rapidity of growth, 

 distance from the ground, and distance from the pith, make more or 

 less difference. 



First, In the more rapidly growing thicker annual layers, the wood 

 fibers are in a larger proportion because the band of large ducts collected 

 at the beginning of the season remains practically uniform, whether the 

 growth be thick or of medium thickness, thus making the increase largely 

 of wood fibers. In woods such as the maples, poplars and willows where 

 the ducts are distributed evenly throughout the season the difference is 

 less marked for the added fibers have with them their proportion of ducts. 



Second, Otlier things being equal a section from the base of the tree 

 will be the toughest and strongest, the medullary rays the thickest, most 

 numerous and if distorted at all the most distorted. The branches 

 farther up become straighter grained, less tough and the annual layers 

 thinner so that conclusions based upon observations taken from a limb 

 might vary considerably from those taken from the base of the same tree. 



Third, The distance from the pith should be noted however for those 

 trees having their pitted ducts collected in bands at the beginning of 

 the season's growth often do not assume their normal position until the 

 third year and after this for some years may be reduced in size and 

 number. Also the medullary rays of all the species examined do not 

 reach their normal size until about one to two inches from the pith, 

 beyond which, disregarding the annual variation of some species, they 

 retain an approximately uniform size. Then to be reasonably sure of 

 a good section for study it ought to be from a known part of a tree six 

 inches or more in diameter, making a moderate growth, at least not 

 struggling for a mere existence, and as nearly a typical specimen as 

 possible. 



The number of species makes necessary some basis for dividing them 

 into groups to simplify their placing. 



A first division naturally falls between the coniferae which have no 

 large pitted ducts and the broad leaved trees which have. A very con- 

 venient general division of the broad leaved trees may be based upon the 

 position of the ducts as follows: 



First, Tlioi-e having a row, rows or bands of distinctly larger ducts col- 

 lected at the beginning of the season's growth. 



Second, Those having the large pitted ducts scattered throughout the 

 year with no tendency to collect in bands at the beginning of the season's 

 growth. 



This may possibly have to be modified somewhat later but thus far it 

 has answered all requirements. 



The elms belong to the first group of the broad leaved trees, and the 

 work on them was extended to include two other members of the same 

 family, the hackberry and mulberry, which I just mention in passing for 

 the remarkable similarity' of their wood to that of the elms. This genus is 

 well marked and quite uniform; and has rapid growing, tough wood, with 

 a yellowish or cream colored sap wood and a brown or greyish brown 

 heart wood. The large ducts are all at the beginning of the season's 

 growth and diminish abruptly to small grouped ones as the season ad- 

 vances. These groups are arranged in nearly (H]uidistant zigzag rows, 

 usually tangling outward toward the close of the season. Often the 



