26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I40 



one such to escape ; in contrast, the Siberian ehn was severely rup- 

 tured in the cambial region down to the ground, even in trunks 8 

 inches in diameter. 



In detail, there is much variation among species, even for the in- 

 tense 1938 frost. This is strikingly true of the Arizona cypress and 

 the several species of yellow pines. Other frosts, of less intensity, 

 show greater variation among different species. 



Among different trees of the same species and among different 

 branches of the same tree, the variations in frost effects duplicate in 

 a large measure what has been said above about species. The case is 

 somewhat different in relation to occurrence in branches; here, the 

 outer 5 to 30 cm. were unaffected by frost. The extent of the frost 

 effects inward on the branch is extremely variable, and most of our 

 sections did not reach inward far enough to reveal their lower limit. 



Where a frost effect is typically developed, it has a characteristic 

 intensity, a characteristic extent around the circuit, and a characteris- 

 tic position in the xylem near, or at the beginning of, the annual in- 

 crement. The frost effects for 1934, for instance, do not commonly 

 extend around the circuit ; they lie either flush against the densewood 

 of the preceding growth layer, or else out one or two cells. In con- 

 trast, the effects of the 1936 frost commonly extend around the circuit 

 and lie nearly flush against the densewood of the preceding growth 

 layer. If the effects are not continuous around the circuit, the local- 

 ized occurrence lies in that part of the increment having the greatest 

 thickness ; in other words, in that region where cambial activity was 

 initiated sufficiently early so that it was caught by the frost. The ef- 

 fects for 1937 are always highly localized as spots which, in many 

 cases, appear to be sunk in the densewood of the first growth layer of 

 the 1937 annual increment. The frost effects for 1938, the most 

 typical and easily recognizable of all, extend completely around the 

 circuit in the majority of instances and are located in the xylem some 

 distance out from the start of the season's growth, a matter of two or 

 three and up to six or eight cells. Further study will bring out the 

 essential characteristics of the 1945 frost ; however, those effects for 

 the years 1939 to 1944, although better known than those for 1945, 

 are less dependable because of greater variation in position, extent, 

 and occurrence. The annual increments for 1932 and 1933 also con- 

 tain frost effects as a characteristic, but so few of our sections contain 

 these increments that it was not possible for us to become as intimately 

 acquainted with them as with the i934-to-i938 sequence. The 

 sporadic frosts, those of 1937, 1939, 1940, 1941, and 1944, have high 

 value for dating purposes when present because they check off the 



