Several laboratory studies have shown that the ingestion of dietary linoleic (18:2 ω6) acid before winter can promote deep and continuous torpor, whereas excess consumption of α-linolenic acid (18:3 ω3) can interfere with an animal's ability to reach and maintain low body temperatures during torpor. As mammalian heterotherms obtain linoleic and α-linolenic acid strictly from the diet, diet selection has been proposed as a mechanism that allows hibernators to ingest levels of linoleic and α-linolenic acid that promote favorable torpor patterns. Here diet, dietary nutrient content and patterns of forage preference of a representative hibernator, the Utah prairie dog Cynomys parvidens, and a facultative heterotherm, the black-tailed prairie dog Cynomys ludovicianus, were examined under natural field conditions. Diets of black-tailed (BTPD) and Utah prairie dogs (UTPD) differed across seasons (BTPD F26,108=9.59, P<0.01; UTPD F38,80=3.25, P<0.01) and elevations (BTPD F26,108=20.15, P<0.01; UTPD F38,80=20.51, P<0.01), and forage preference indices indicate that neither species randomly selected plant species relative to their abundance on colonies in any season. Black-tailed prairie dogs did not consume or avoid consumption of plant species based on levels of total lipids, linoleic acid, α-linolenic acid or nitrogen. Considering only the plants consumed, black-tailed prairie dogs appeared to prefer plants with low levels of α-linolenic acid (F1,19=5.81, P=0.03), but there were no detectable relationships between preference and other nutrients. Utah prairie dogs consumed plants higher in α-linolenic acid (t=1.98, P=0.05) and avoided plants high in linoleic acid (t=−2.02, P=0.04), but consumption-avoidance decisions did not appear to be related to nitrogen or total lipids. Of the plants consumed, Utah prairie dogs again preferred plants high in α-linolenic acid (F1,17=4.62, P=0.05). Levels of linoleic and α-linolenic acid were positively correlated in plants consumed by prairie dogs (BTPD Pearson r=0.66, P<0.01; UTPD Pearson r=0.79, P<0.01), reducing the opportunity for independent selection of either lipid.