This study was conducted at the east of Karnali River in Bardia National Park in hot-dry season on the month of May, for the analysis of nutrients of important food tree species of Asian elephant (Elephas maximus). Survey block was first determined in the topographic map and the first transect line was selected randomly west to east, inside the survey block. Other transects were drawn parallel to the first transect line. Transects were 1 km. apart from each other and each transect was 3 to 3.5 km. in length. Samples of bark and leaves were collected from the main trunk and branches of the trees which were fed by wild elephants. Samples were collected within a circular sampling plot (n = 74) of radius 15 m. at every 500 m. interval of each transect lines (10 transects and total length = 32 km.). Samples of bark and leaves were collected from three different heights (where available); i.e. bottom (< 2 m.), mid (2 – 4 m.) and top (> 4 m.) of the tree. Each sample was collected in separate plastic bags with appropriate tags.

Estimation of the chemical constituents of the bark and leaves were done by following the procedure of A.O.A.C. (Association of Official Analytical Chemist) - 1975, and the method given by Talapatra, Ray and Sen (1940). In the present work, investigations were made on estimation of: dry matter, nitrogen, crude protein, ether extract or fat, crude fiber, ash, calcium, phosphorus, nitrogen free extract or carbohydrate, organic matter, total carbohydrate and sodium.

Statistical analyses were done by using the statistical software “Sigma-stat”. t-test was applied to assess the nutritional value between bark and leaves of the food trees. Mann-Whitney Rank Sum Test was applied where normality test and equal variance test had failed. One-way analysis of variance (ANOVA) was used to assess the nutritional value within food trees between bark and leaves according to the tree height. Kruskal-Wallis one-way ANOVA on Ranks was used wherever normality test failed.

Major food tree species found to be eaten by the elephants were Sindure (Mallotus phillippinensis), Sadan (Desmodium oojeinense) and Sissoo (Dalbergia sissoo). Investigations revealed that the most preferred food tree species was M. phillippinensis (61.54%) followed by D. oojeinense (30.77%) and D. sissoo (7.69%).

Bark of M. phillippinensis contained higher nutrients (Dry Matter and Crude Fiber) and minerals (Calcium, Phosphorus and Sodium) and has greater potential for elephant feeding, especially in the hot-dry season in Bardia National Park, when the grasses have poor nutritive value. This is the main reason behind consuming the bark of M. phillippinensis than the leaves in order to fulfill their mineral deficiencies during the hot-dry season. There was not much differences in the content of minerals in the bark and leaves of D. oojeinense and D. sissoo.

The concentration of phosphorus was found to be lesser at the bottom of M. phillippinensis in comparison to mid and top part, and in comparison to other two food tree species (D. oojeinense and D. sissoo). Phosphorus is often a limiting nutrient for herbivores, so, probably this could be the reason for the preference of browsing above the bottom part of M. phillippinensis.