180 JOURNAL OF FORESTRY 



plant or group of plants. Therefore, to be reliable, a key which uses 

 this diagnostic criterion must be based upon a wide range of specimens. 



Type and Arrangement of Vessels 



The structure and arrangement of vessels, or conducting passage- 

 ways, have been used extensively as diagnostic criteria in the classifi- 

 cation and identification of wood. For example, the patterns, made 

 by this grouping of the vessels, have been used as a means of dis- 

 tinguishing the wood of various species of Ulmus. In the red or slip- 

 pery elm, Ulmus fulva Michx., there are supposed to be several rows 

 of large vessels in the springwood ; in the winged elm, U. alata Michx., 

 on the contrary, there is supposed to be only a single, row of large 

 vessels in this portion of the growth rings. In the summerwood of 

 the former species, the vessels are said to occur in diagonal or flame- 

 like clusters ; but to form tangential or horizontal bands in the winged 

 elm. 



That such criteria may be subject to considerable variation, even 

 within a single species, is indicated in figures 3 and 6. Figure 3 illus- 

 trates a cross-section of the wood, taken from a "butt" log of the 

 common white elm, U. americana L. Figure 4 shows the structure of 

 the wood in a "top" log of the same tree. As the photomicrographs 

 were made at the same magnification, it is obvious that the annual 

 ring was of the same width in both specimens. 



In the genus Prunus, the structure and arrangement of the vessels 

 fluctuates much as does the distribution of parenchyma. Figures 10 

 and 15, cross sections of the wood of Prunus serotina Ehrh, illustrate 

 the type of vessel structure that occurs commonly in the northern 

 representatives of the genus. In such tropical species a^ P. sphaero- 

 carpa, P. janvanica, etc., the vessels are usually less numerous, larger 

 and thicker walled, and show a greater tendency to occur in short 

 radial rows (figs. 11 and 12). In many of the sclerophyllous types, 

 such as P. ilicifolio and P. integrifoHa (fig. 9), the vessels are small, 

 very thick-walled and aggregated into larger radial clusters. 



It is evident, accordingly, in view of the fact that there is no existing 

 method for accurately estimating their limits of variability in any 

 particular species, that the structure and arrangement of the vessels 

 cannot be used safely as diagnostic criteria in keys that are based upon 

 a limited number of specimens. 



