PRODUCT STABILITY--PART II 321 resulted from which a viscosity half-life could be calculated. This half- life decreased with increasing molecular weight. In a study of the viscosity of suspending agents, Joslin and Sperandio (49) noted straight line (and some curved) relationships obtained by plotting viscosity rs. time. Not surprisingly, some experiments resulted in a family of lines with respect to different storage temperatures and also, some samples became thicker with time, some thinner. Storage at elevated temperatures accelerated the particular changes. Zacek (50) illustrated the use of ductility as a response variable of pharmaceutical suspensions. In essence, he suggests placing the suspen- sion between two plates which are then pulled apart vertically. A thread between the plates forms, and the distance at which it breaks is measured. With respect to stressing suspensions, the freeze-thaw cycle technique is very applicable. This treatment facilitates particle growth and may indicate the probable future state of affairs after long storage at room temperature. Schwarz and Levy (51) studied the viscosity of sodium alginate solu- tions after freezing and thawing. They showed that the viscosity in- creased upon such treatment. They also demonstrated that excessively long shearing lowered the viscosity, although the freeze-thaw cycle caused all to have the same viscosity regardless of the initial. They also pointed out, however, that an increased rate of shearing may give a permanent loss of viscosity, as will higher temperatures, because of the degradation of polymer chains. Head and Lauter (52) used an ultrasonic generator to study the depolymerization of natural. polymers. It is known that such insonation will depolymerize, but it also can facilitate gum hydration. They showed through viscosity measurements that the molecular weight de- creased and that a straight line plot resulted when the molecular weight was plotted against the duration of irradiation. Carrageenan, agar, lo- cust bean gum, and methylcellulose all degraded first order gum tragacanth degraded according to zero order kinetics karaya and acacia were not degradable by their method. Semi-solids and Solids Urbanyi and co-workers (53) measured the changes in reflectance of colored tablets with a reflectometer attachment on a spectrophotometer, choosing the best wavelength for observation. Reflectance was com- pared to a standard. They used this technique to determine the in-

322 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS fluenee of light of 45 and 550 foot-candles intensity. Absorbance was plotted against time, and some first order rate constants were obtained, although the plots were not ideal as the lines were segmented. Natu- rally, the constants increased at the higher light intensities. Lachman et al. (54) in additional work which was part of an extensive study of tablet dyes checked the fading of the corresponding lakes. From absorbance and time data, they obtained first order rate constants and showed that the lakes studied had poorer photostability than the dyes. Carstensen and collaborators (55) employed photometric reflectance to observe the appearance of tablets and powders stored at elevated tem- peratures. Their data provided them with both zero order and first order rate constants which were amenable to Arrhenius plot treat- ment. Everhard and Goodhart (56) studied the fading of dyes in tablets. Their aim was to quantitate the time-intensity relationships involved in the storage of a single tablet dye at various concentrations (0.015 to 0.060%). The tablets were stored under 655 foot-candles, fluorescent lights at 80 foot-candles, fluorescent light and amber glass at 11 foot- candles, and under incandescent light of 50 foot-candles. They pointed out that others had plotted log of log i/R vs. time where R represented the fraction of light reflected at a specified wavelength. These workers pointed out that the previous work inferred that log i/R was propor- tional to concentration but that this was not so because different rate constants were obtained for different dye concentrations. Thus, they noted the need for an equation to relate reflectance to concentration and also a need to bring light intensity into the kinetic picture. They showed that a parameter called t•t was proportional to concentration where: v•t = (1 -- Rt)'•/2Rt and Rt = the measured reflectance at the minimum wavelength in a reflectance-wavelength curve. Then, con- sidering fading to be proportional to time and intensity, a straight line resulted when •t was plotted vs. I X t in foot-candle hours. McKeehah and Christian (57) studied the color stability of a benton- ire base cream with an integrating sphere reflectometer, which is useful for the study of opaque solids and semisolids. When the creams were stored at various elevated temperatures, they were able to plot reflect- ance (relative to magnesium carbonate) vs. time and obtain straight lines. They were also able to construct Arrhenius plots to permit predic- tion of the time when the color would be unsatisfactory. They also proposed use of the technique for the study of coated tablets, powders, and granulations.