222 REPORTS ON INVESTIGATIONS AND PROJECTS. 



applied (and removed) at intervals of loo grams, from loo grams 

 to 1 200 grams, inclusive. It was found, as is well known, that 

 the strain due to any particular load was dependent upon time. If 

 the strain is plotted as ordinate and the time as abscissa, the time- 

 effect curve passes through the origin, is concave to the axis of time, 

 and ultimately becomes asymptotic to a line parallel to this axis. A 

 unique solution of our problem would consist in finding the relation 

 between the successive ordinates to these asymptotes, as their ordi- 

 nates represent the amount of strain which takes place in an infi- 

 nite time. This work has not yet been completed, but a thorough 

 investigation of the curves is being made with good hope of deter- 

 mining their complete analytical representation. For the after- 

 effect, or the time- effect on wires newly released from stress, I have 

 found a satisfactory analytical expression. 



At present, steel tapes (cross-section ^ inch by 0.007 inch) are 

 being used instead of copper wire. The obvious advantages are that 

 the tapes are more nearly homogeneous, they do not require to be 

 overstrained in order to be brought into the elastic state, and, finally, 

 steel is probably the material which corresponds more nearly to that 

 of the interior of the earth. A series of experiments was begun on 

 August 10. Successive loads at intervals of 10 kilograms are being 

 applied, and the time-effect curves determined as before. The attempt 

 is being made to continue the observations until the strain becomes 

 almost independent of the time, thus diminishing the error due to 

 extrapolation. The total increment in length due to a load of 20 

 kilograms is 13.7 cm. The time-effect is 0.07 cm. It seems to be 

 possible to attain the degree of accuracy in measurement that was 

 anticipated, namely, o.oi mm., or a strain of i in 15 x 10*. This is 

 approximately the same order of magnitude as obtained by the use of 

 interference methods. The time-effects tend to delay the progress of 

 the work in so far as the determination of the stress- strain relation 

 is concerned, but information in regard to these phenomena alone is 

 very valuable. The labor and expense of continuing the experiments 

 will be comparatively small, as in general only two or three observa- 

 tions a day are necessary, and these can be made without materially 

 interfering with the progress of other work. 



Attempts have been made to determine a stress-strain relation for 

 some anomalous substances, such as catgut and a pseudo-solid 

 composed of pure albumen and sugar. The experimental part of this 

 work is almost completed. A few preliminary experiments have 

 been made on crystals with interesting results. 



