Investigations of botanical composition of diets must consider sources of variation in diet composition. Annual cycles generally exist in diet selection and are controlled by the availability of alternative forages and toxin concentration of the plant. Short-term cycles result because of postingestive consequences of plant toxins. Finally, changes in weather may cause episodic variation in the diet composition. These sources of variation can significantly affect attempts to estimate diet composition. An experiment was conducted on the Texas A&M University Research Station at Sonora to evaluate seasonal patterns in juniper consumption and identify optimum sampling regimens. Twelve non-pregnant Angora female goats were categorized as high or low juniper consumers based on fecal near-infrared reflectance spectroscopy (FNIRS) predictions from samples collected in Sept 2003 and Jan, Mar, and Jun 2004. These animals were continuously grazed on a 16 ha pasture supplied with free choice mineral and water from October 2004 through the end of September 2006 and fecal samples for FNIRS were collected two times per week. Juniper levels were tested for annual and short-term periodicities as well as episodic differences caused by weather events. Percent juniper was positively related to 30 d moving average precipitation and negatively related to 30 d moving average temperature, which accounted for 17% of the annual variation in diet selection (P<0.001). Climatic variable that would indicate episodic response to climate were not identified. One and 2 d lags of percent juniper were also significant (P<0.001), which are interpreted to indicate short term cyclic response to phyto-toxins and accounted for an additional 54% of the variation in diet selection. Fourier series analysis in combination with the 2 significant autoregressive variables indicate an approximately 21 d peak to peak periodicity in consumption of juniper.