sample is connected to the chart movement amplifier controlling the Y component in a standard manner.The amplifier circuit in the pen servo circuit is sufficiently sensitive that it can easily be overloaded into mechanical oscillation. The potential applied to the photocell bridge is reduced to about 10 mv. by a drop resistor to eliminate this overdrive and subsequent hunting. Stray light must be reduced with appropriate baffles and shields ; however, complete darkness is not required.A typical expansion curve obtained from this revised apparatus is shown in Figure 2. This curve is for a sample of uncured polybutadiene, about 98% in the cis-I ,4 configuration. The sample in the expansion cell was precooled with liquid air, beginning the run. Warmup from about -8 mv. The melting point, T,, is evident a t -0.2 mv. (-4"C.), above which linear expansion again can be noted. Frozen stresses account for the apparent shrinkage just above the melting point.While the work reported thus far in this note and in reference 1 has covered only the temperature range below room temperature, the range above room temperature can bc studied easily by replacing the Dewar system with a small oven, heated a t a controlled rate. Studies typical of this latter kind are illustrated by glassy changes in polyvinyl chloride a t about +85"C. and by crystalline melting points in polyethylenes about 130°C. Most polymers studied a t higher temperatures are too low in molecular weight, to show simple rubbery expansion above either l', or l',,,. Instead, at the higher temperatures, they generally creep or flow rapidly, apparently indicating contraction. Cooling curves from very high temperatures can be obtained if the mirror tripod is properly adjusted or compensated for the expected creep.
Gravimetric Dilution ViscometryFrequently it is desired to estimate molecular weights of very small samples of polymer, e.g., a few milligrams. If viscometry is to be used, dilution viscometers requiring only about 1 ml. of polymer solution are desirable. At least two such viscometers have been reported. One was described by Hart,' and the other, which we have used in this laboratory, is cammercially available.2 Such viscometers require the delivery of accurately known portions of solution and solvent. Volumetric methods require careful calibration of volumetric flasks and of pipets for both solutions and solvents, since drainage errors and reproducibility present serious difficulties. Even so, the precision of calibration is not as good as might be desired, particularly with volumetric flasks.These problems can be overcome simply by using a purely gravimetric procedure. The initial polymer concentration is determined gravimetrically on a weight per cent basis. The polymer solution and successive solvent portions are introduced to the viscometer by means of a syringe which is weighed on a semiautomatic, single-pan, analytical balance before and after addition of liquid to the viscometer. We have used B-D Luer-Lok Multifit syringes (registered t r a d e marks o...