Our study area was situated in the south-western part of the Bardia National Park in lowland Nepal (ca. 100 m.a.s.l.), bordered by human settlements in the east and the Geruwa River in the west. The climate was subtropical monsoonal with heavy rain fall from July to October. The vegetation consisted mainly of Sal forests Shorea robusta and patches of grasslands dominated by Imperata cylindrica in the east. A patchwork of Saccharum spp. grasslands and a variety of successional forest types dominated the vegetation along the river. Among the ungulates in the park, chital and hog deer were the most abundant; muntjac, barasingha, sambar, nilgai and wild boar occurred in lower numbers. Tiger was the most abundant carnivore; striped hyena and dhole were less common.
Three Leopards, one adult female and two adult males, were captured in box traps and radio-collared during 19992001. The female (F1) and one of the males (M1) were monitored for a period of approximately two years, whereas the other male (M2) was followed for 3 months. We used two radio monitoring programs: (I) 24-hour intensive tracking sessions (F1 = 13 days, M1 = 4 days, M2 = 2 days), with precise radio fixes every hour, and (II) daily radio tracking for periods of minimum 10 days.
The first five sessions of 24-hour intensive tracking of F1 were done on consecutive days, whereas the following sessions were not.
F1 was tracked daily for 180 days during eight periods in the dry seasons of two consecutive years (October 1999 March 2000 and January 2001 March 2001). She gave birth to two cubs in June 1999 and two cubs in August 2000, and was therefore with dependent young aged four to nine months during the two main sampling periods. We allowed occasional gaps of a maximum of one day in the daily monitoring plans. However, such interruptions only occurred in 9% of the days included in the analyses.
The movement pattern of Leopards provided cues that we used to find kills. We searched for kills within areas where the Leopards had been active for several consecutive hours (during the 24-hr intensive tracking sessions), or for at least two consecutive days (during the daily tracking periods). Radio fixes were usually obtained between 9 am and 5 pm during the daily radio tracking. When a radio fix was positioned less than ca 500 m from the location of the previous day, we tracked the Leopard in intervals of ca 1 hour during the rest of the day. This procedure was followed to make sure that we obtained fixes while the Leopard was feeding at the kill. We were careful not to disturb the Leopards while they were feeding to avoid that kills were abandoned. In order to estimate the biomass consumed from each kill, we multiplied the proportion consumed by the live weight of the kill (LW). We assumed that the percentage of wastage from completely consumed kills in this study was
LW < 5 kg = 0%, LW 5-25 kg = 5% and LW > 25 kg = 30%. Deviations from the maximum amount consumed were estimated visually.
During the 13 days of intensive tracking of F1 we found 3 Chital, 1 Peacock and 1 Paddy bird. M1 killed 2 adult male chital and 1 domestic dog during 4 days of intensive tracking. No kills were recorded during the 2 days of intensive tracking of M2. All edible biomass was consumed from the kills made by F1, while two of the kills of M1 were only partially eaten. We estimated that the three Leopards consumed 89.2 kg of meat during the 19 days of intensive tracking, rendering an average daily consumption of 4.7 ± 0.3 kg of meat during this period for all the Leopards. The males consumed 4.3 ± 1.1 kg/day and the female consumed 4.9 ± 0.4 kg/day. In three instances, the Leopards fed on kills that had been made before the tracking sessions started. Hence, since only five of the eight prey items were killed during the tracking periods, the sample size was too limited for estimating kill rates.
We found 25 kills during the 180 days that F1 was tracked daily. Among these, there were 14 Chital (6 adult males, 4 adult females, 1 yearling, 2 fawns, 3 of unknown sex and age), 1 Indian muntjac, 4 primates (2 rhesus macaques and 2 common langurs), and 6 birds (4 peafowl, 1 unidentified raptor and 1 paddy bird). The utilizable body mass of all but 3 kills was consumed entirely.
There was a significant and positive relationship between the live weight of the kills and the number of days spent at the kill site.
However, we found a significant relationship between daily consumption rate while feeding at kills and age of cubs, indicating that kills were consumed faster when the age of the cubs increased.
The daily straight-line movement distances during periods of confirmed kills averaged 194 ±43 m, and 1128 ± 100 m between kills. By examining the distribution of the radio fixes of F1, we identified 9 periods with markedly short daily movements (131 ±20 m), indicating that kills had been consumed that we were unable to locate. In order to approximate the weights and species categories (ungulates and other prey) of the suspected kills, we analyzed the relationships between the total amounts of meat consumed from the confirmed kills, the number of days spent at the kill site and the age of the Leopard's cubs.
The predicted amounts consumed from six of the suspected kills were 9.4 kg, and ranged between 27 and 52 kg in the three remaining kills. Hence, a minimum of 3 suspected kills were ungulates, as these were the only available prey species weighing more than 27 kg. Since nearly all confirmed ungulate kills were completely consumed, we assumed that the six suspected kills with consumption < 10 kg were not ungulates. The estimated kill rate of ungulates by F1 based only on confirmed kills was 1 per 12.9 ± 1.0 days. Including suspected kills produced an estimate of 1 per 10.6 ± 0.7 days. Including all kill categories, the kill rates were 1 per 7.8 ± 0.6 days and 1 per 5.6 ± 0.4 days among confirmed kills and confirmed + suspected kills, respectively. The female F1 consumed up to 4.0 ± 0.9 kg of meat daily.