centage of error greater than ten, and the average percent of error for 

 the whole group is 10.1 or one mistake for every ten specimens examined. 

 Considering that 7748 pieces of wood were examined by twenty-seven dif- 

 ferent persons, the percentage of error for the entire group should not be 

 far from that made by the average person accustomed to handling these 

 woods. 



Of the ring-porous woods only two catalpa and black locust have 

 more than ten per cent error. The excess above that figure appears to be 

 due in part to a common mistake made in one set of papers where catalpa 

 was incorrectly identified as butternut. This is more of a blunder than 

 a mistake, for butternut is not only classified under a different group but 

 the color of the two are quite dissimilar. It is only reasonable to suspect 

 that this error was due to a "doctored" specimen for by eliminating the 

 twelve mistakes the total percent of error is reduced to 11.6, which is more 

 consistent with the expected results. 



Although the total number of species having a percentage of error 

 greater than ten is the same in the diffuse as the non-porous group, the 

 average percentage of error for the former is four-fifths that of the latter. 

 This is due no doubt to the fewer number of specimens examined. In the 

 case of the Spanish cedar the figures are misleading, for there is a close 

 resemblance between it and West Coast and Honduras mahogany. It could 

 be mistaken for either one of these woods, especially if the characteristic 

 odor were not very prominent. 



The extremely high percentage .of error for teak is also very mislead- 

 ing to those who are not acquainted with it and Oregon myrtle, for which 

 it was mistaken. Although the two are very close together in the order 

 of classification, their general appearance is very dissimilar, and to mis- 

 take one for the other would be considered a gross blunder on the part 

 of even a novice. As an added proof that Oregon myrtle is not readily 

 mistaken for other woods, a glance at the table will show that out or 

 the forty-one times that the wood was examined not once was it identified 

 incorrectly. By referring to the column below the name at the top of the 

 table, it will be seen that only once was any other wood mistaken for 

 Oregon myrtle, and that was in the ease of the blunder referred to above. 



On the whole the last group does not show such conclusive results as 

 might be expected from the 4382 specimens examined, due, no doubt, to 

 the unequal representation of species among the samples submitted. If 

 all species that are represented by less than 100 specimens are left out, 

 the average percentage of error will be materially reduced and more nearly 

 equal to that of the ring-porous group. 



Conclusions. 



Considering only the figures shown in the table, one may conclude 

 that such non-porous woods as hard pine, eastern larch, Douglas fir, red 

 spruce, western hemlock, true fir and Port Orford cedar are the commercial 

 woods most likely to be incorrectly identified. To conclude that western 

 yew is easily identified is correct, and the mistake marked against it is 

 hardly excusable owing to the dissimilarity of texture between it and 

 redwood. 



The tabulated results would lead one to conclude rightly that the 

 ring-porous woods are the easiest to identify. It might be said that there 

 is less similarity between the woods of this group than between those of 

 the other two, due to the arrangement and size of pores and the charac- 

 teristic colors. 



No satisfactory conclusion can be drawn from the table for diffuse- 

 porous woods, but from the experience of the writers it was found that the 

 mahoganies; maple, birch, beech and alder; basswood, cottonwood, and 

 tulip poplar, were the most difficult to identify correctly. 



The gross blunders appear to be equally distributed among the three 

 groups, in each case being only a very small percent of the total mistakes. 

 In working up the data it was found that the papers for the year 1915-16 



42 



