400 CARNEGIE INSTITUTION OF AVASHINGTON. 



whereas it was originally assumed that but 2 constants were desirable. 

 Similarly, it was at first assumed that one might express the wind- 

 effects sufficiently well in terms of the mean velocity and prevailing 

 direction of the wind for each 24-hour period. The investigation to 

 date indicates that it is necessary to take into account the wind velocity 

 and direction for each hour, and that possibly at least 8 constants are 

 necessary to express the effects adequately, instead of 4 constants as 

 at first assumed. 



The data, including nearly all necessary meteorological observations, 

 are in hand for the equivalent of 223 months of observations of the lake 

 surface at one gage. Of these 223 months of observations, but 107 

 have yet been used in the computations, and even for these months it 

 is now known that considerable improvements are possible in the 

 method of computation thus far used. Hence, much more progress is 

 possible in the investigation. It is now reasonably certain that the 

 evaporation under prevailing conditions will be determined with a 

 probable error not greater than one-fifth of the evaporation. 



No serious attempt has thus far been made to study stream-flow, as 

 that portion of the investigation is intended to be based on more infor- 

 mation in regard to evaporation than is now available. 



Howe, Henry M., Columbia University, New York, New York. Investiga- 

 tion into the physics of the iron carbon alloys. (For previous reports see 

 Year Books Nos. 6-13.) 



During the year 1914-15, most of the work has been done on the 

 nature of the deformation hues in metals, and especially in alpha iron 

 and gamma iron. The results will be published in a volume on the 

 metallurgy of steel, about to go to press. 



The following is a brief resume of the lines of my investigations, with 

 a statement of some of the facts educed : 



1 . The details of the appearance of slip bands in alpha and gamma 

 iron and in copper, of Neumann bands in alpha iron, and of annealing 

 twins. 



2. The retardation of the Brinell hardness test greatly increases the 

 apparent hardness of iron by giving an opportunity for the trans- 

 formation incited by the earlier stages of the test to take effect, and to 

 raise the elastic limit before the later stages occur. 



3. The path of rupture in pure alpha iron and in mixtures of alpha 

 iron and pearlite. 



4. While the deformation in the early stages is in appreciable part 

 intergranular — that is, between the polyhedral allotriomorphic crystals 

 of which the metal is composed, commonly called grains — yet the path 

 of rupture passes by preference as far as possible from the grain 

 boundaries and therefore along grain centers. I have sought methods 

 of explaining this phenomenon. The present evidence indicates that. 



