compound system of levers, the various errors and difficul- ties are likely to be magnified in the same proportion. The principal difficulties met with were due : (1) to the weight of the compound lever which exerted a great tension on the growing plant, (2) to the yielding of flexible connec- tions by which the plant was attached to the first lever, and the first lever to the second, and (3) to ihe friction at the fulcrums.
Weight of the Lever. — As the first lever is to exert a pull on the second, it has to be made rigid. The second lever serves as an index, and can therefore be made of fine glass fibre. The securing of rigidity of the first lever entails large cross section and consequent weight, which exerts considerable tension on the plant. Excessive tension greatly modifies growth ; even the weight of the index used in self-recording auxanometers is found to modify the nor- mal rate of growth. The weight of the levers introduces an additional difficulty in the increased friction at the fulcrums, on account of which there is an obstruction of the free movement of the recording arm of the lever. The conditions essential for overcoming these difficulties there- fore are : (1) construction of a very light lever possessing sufficient rigidity, and (2) arranging the levers in such a way that the tension on the plant may be reduced to any extent, or even eliminated.
I found in navaldiim, an alloy of aluminium, a light material possessing sufficient rigidity. The first lever is con- structed out of a thin narrow sheet 25 cm. in length ; it has, as explained before, to be fairly rigid in order to exert a pull on the second without undergoing any bend- ing ; this rigidity is secured by giving the thin narrow plate of the lever a T-shape. The first lever balances, to a certain extent, the second. Finer adjustments are made by means of an adjustable counterpoise B, at the end of the levers. By this means the tension on the plant can
