190 Mr. A. Mallock. Groivth of Trees, with a Note on 



mounted on the stiff support E, and the edge F is kept pressed against the 

 face of the prism by a light spring S.?, which also tends to turn the plate 

 about the edge (and the knife-edges K in line with it) outward from the face 

 of the prism. The distance between E' and B' (fig. 3), and therefore the 

 angle between the plate and prism, is limited and defined by the two studs 

 on the rockers, thus any change in the girth of the tree causes a corre- 

 sponding change in the angle between the two glass surfaces. 



When preparing for a set of observations, the bedplate is first secured to 

 the tree, and the " invar " tape is then passed round the smoothed track on 

 the bark and over the rockers, and the tension springs are hooked on to the 

 tape and secured. The prism holder is next put in place, and the two 

 rockers are turned until the studs bring the surfaces of the plate and prism 

 into contact. The interference bands are now very broad (in fact, if the 

 surfaces were truly flat and truly in contact, ' the field would appear of one 

 uniform shade). The micrometer screw is now turned until the bands 

 assume a width convenient for observation. 



In order to define the position of the bauds, a narrow central streak M 

 (fig. 4) is painted on the face of the prism with an alcoholic solution of 

 safranin or other suitable anilin colour. This, when dry, is 

 only a small fraction of a wave-length in thickness, but the 

 colour is quite apparent by transmitted light. A narrow 

 gap T is made in the streak by means of a pointed piece 

 of hard wood wetted with alcohol, and this serves as a mark 

 to which the position of the interference bands can be 

 referred. When very small variations of the girth of the 

 tree, such as may occur every few minutes, are to be observed, 









T 





M 



Fig. 4 



it is convenient to measure them by estimating the fraction 

 of the band which crosses the gap, but, for ordinary work, where the intervals 

 between the observations are an hour or more, the plan adopted is to bring 

 back the bands to their previous position by turning the micrometer screw, 

 whose readings give directly the variation of angle between the plate and 

 prism which has taken place since the previous observation. Here the optical 

 theory does not enter, and the bands are simply used as delicate callipers. 



To facihtate observation, a second right-angled prism (P', fig. 3) is 

 cemented to the interference prism so as to reflect the emerging pencils in a 

 horizontal direction to the small telescope, seen in fig. 1, whose aperture is 

 reduced sufficiently to give good definition to the bands. 



The first trials of the apparatus were made in April and May, 1917, at 

 Kew, where, by the kindness of Sir David Prain, I was enabled to make 

 observations on several species of trees. It was soon found that the rate of 



