Analysis of Bamboo Preferences

Analysis of bamboo preferences in a pair of captive giant pandas (Ailuropoda melanoleuca)

Loraine R. Tarou, Jessamine Williams, David M. Powell, Raquel Tabet, and Mary Allen

Optimal foraging theory postulates that animals forage in such a way as to maximize the energetic gains and minimize the energetic costs of foraging. It has been argued that the nutritional quality of a food source is more important for herbivores than for carnivores or insectivores because plant material is generally lower in quality and harder to process and glean nutrients from than other types of food. Therefore, to meet their nutritional requirements, herbivores need to be more selective in their choice of diet than other animals. The giant panda is somewhat of an anomaly in the animal kingdom in that it is a carnivore that subsists almost exclusively on an herbivorous diet. Bamboo comprises 99% of their diet (Schaller, Jinchu, Wenshi, & Jing, 1985). Pandas have several morphological adaptations for the efficient processing of bamboo including an enlarged radial sesamoid bone, often called the pseudothumb, which is used for grasping bamboo culms between the bone and the palm of the hand, flattened molars with increased surface area for grinding bamboo culms, and extensive jaw muscles that move their teeth from side to side to break down the tough plant material. However, their digestive systems do not appear to be specialized for processing bamboo. Rather than having a complicated stomach that is generally characteristic of an herbivore, they have a relatively simple stomach similar to that of the typical carnivore. Because of this type of simple digestive system, pandas are only able to digest approximately 12-23% of the bamboo they consume. Therefore, pandas must eat a large amount of bamboo throughout the day and night to gain enough nutrients to survive. This should make them selective for the type and parts of bamboo that are either the most nutritious or easiest to process.

Pandas have been reported to feed on as many as 42 species (Taylor, 1993), but in some areas may specialize on as few as two (Schaller and colleagues, 1985; Taylor & Qin, 1987). Pandas selectively feed on plant parts of higher nutritional value (leaves, young shoots). Nutritional research indicates that the leaves of bamboo are more digestible than culms or branches, because the leaves contain higher crude protein levels (13%) than other plant parts (3%) Leaves contain the highest levels of crude protein, which is 90% digestible. In comparison, the cell wall, cellulose and lignin, are no more than 10% digestible. Given the low nutritional quality and low digestibility of bamboo, it is likely that pandas will demonstrate a preference for the species of bamboo that provide the greatest nutritional value and energy intake per quantity consumed. The purpose of our study was to determine whether two captive giant pandas exhibit preferences among three bamboo species. We hypothesized that the pandas would prefer species with larger leaves and culms that are smaller in diameter than other species.

Experimental Procedure:

The pandas included in our study were 1 male, Tian Tian, and 1 female, Mei Xiang, housed at the Smithsonian National Zoological Park. Their normal diet consisted of 15-20 kg of yellow groove bamboo (Phyllostachys aureosulcata), leafeater biscuits which are high in fiber, and produce including apples, carrots and sweet potato. The pandas at the National Zoological Park are housed together outdoors from 8 am – 5 pm and are separated indoors from 5 pm – 8 am.

To determine bamboo preferences, we conducted classic two-choice preference trials using three species of bamboo, Phyllostachys bissetii (bissetti), Phyllostachys nigra (black bamboo), and Pseudosasa japonica (arrow). The bissettii and black were similar to each other in leaf and culm size, whereas the arrow bamboo had significantly smaller leaves and culms than the other two species. All three of these species were novel to the pandas and were not part of their regular diet. In each trial, the bamboo species were randomly paired and placed in one of two predetermined locations in one of two indoor rooms to control for possible position effects resulting from asymmetry of the enclosure. Mei Xiang was provided with 5 kg of each bamboo species, and Tian Tian was offered 8 kg. This was twice the total amount of bamboo normally provided in the evening meal so that if there was a clear preference for one species over another, the pandas could feed to satiety on the preferred species.

A total of 18 choice trials were conducted every other day at approximately 5:00 in the afternoon. The pandas were observed for a single, one-hour session immediately following exposure to the bamboo at approximately 1700. Behaviors were recorded using continuous focal sampling with The Observer and included the amount of time spent feeding on each species, the amount of time spent eating different parts of the plant and other behaviors. Plant parts included the culm, leaves and a combination of leaves and branches. Additionally, total weights of bamboo offered at night and left over in the morning were recorded in a manner similar to that used in the pilot study.

To account for bamboo water loss the dry matter concentration of each bamboo species was calculated and used as a correction factor in the bamboo consumption data. To determine dry matter concentration of the bamboo, nutritionists collected approximately 10% of the fresh bamboo offered on six trial days and 10% of the unconsumed bamboo leftover the morning after six trial days. After collection, samples were stored in plastic bags in a walk-in refrigerator until processed. Samples were separated into leaves, branches and culms, which were stored in plastic bags, weighed and stored again in the cooler until removal for dry matter determination.

A Forced Air Draft Oven set at 50C was used to dry the processed bamboo. Samples were reweighed daily until they reached constant weight, usually after 3 days. Once dried, the samples were stored in plastic bags at room temperature. Samples were ground in a Wiley Mill using a 1-mm screen, weighed and put in a drying oven for 3 days. Finally, samples were removed from the oven, put in a desiccator for one hour to cool and weighed.

Dry matter concentrations of the samples were calculated according to the following formula:

Dry matter concentration = (DM culm) + (DM branch) + (DM leaf)

(FW culm) (FW branch) (FW leaf) = [(DM culm)+(DM stem)+(DM leaf)]/[(FW culm)+(FW stem)+(FW leaf)]*100

where DM represents the dry matter weight for each plant part and FW represents the fresh weight for each plant part. Since dry matter concentration was not determined on bamboo offered and refused in all trials, average dry matter determined in selected trials was applied to all trials.

Dry matter of bamboo consumed during each trial was calculated according to the following formula:

kg consumed = (kg FW offered x DM offered) – (kg FW refused x DM refused)

where FW represents the fresh weight of bamboo offered and refused and DM represents the species specific dry matter concentration for the bamboo offered and refused.

Findings

As predicted, Mei Xiang showed a strong preference for the leaves of all three species of bamboo presented to her. Tian Tian showed a strong preference for the leaves of both black bamboo and bissetii bamboo, but showed an equal preference for the leaves and culm of arrow bamboo. Arrow bamboo has a culm that is smaller in diameter than that of the other species tested.

In terms of the total amount of time each panda spent feeding on each species during the one hour data collection session, both Mei Xiang and Tian Tian exhibited a strong preference for arrow bamboo, spending more time eating arrow bamboo when it was offered than they spent eating the other species. In terms of overall consumption over the course of the night, Tian Tian still exhibited a strong preference for arrow bamboo and Mei Xiang consumed similar amounts of both arrow and bissetii.

Further indicating that arrow was the preferred species of bamboo was the fact that the position of the bamboo affected behavior only when bissetii and black bamboo were offered with each other during the trial. It was discovered that both Tian Tian and Mei Xiang had favorite eating positions in the room. Tian Tian’s favorite position was that position that was closest to his favorite sleeping position. Mei Xiang’s was the position that was closest to the entrance of her indoor holding facility. When the choice was bissetii vs. black bamboo, both pandas would spend more time eating the species that was in their favorite position in the room. However, when arrow was offered as a choice with either black or bissetii, both pandas would preferentially eat arrow regardless of its position in the room. That is, preference for arrow took precedence over their preference for position.

In conclusion, leaves were the most preferred part of all species (except arrow for Tian Tian). Both pandas showed behavioral preferences for different species of bamboo such that arrow was preferred over bissetii bamboo which was preferred over black bamboo. There were individual differences in overnight consumption such that Mei Xiang consumed equal amounts of arrow and bissetii, whereas Tian Tian exhibited a clear preference for arrow. The position of the bamboo was important and affected behavior only with less preferred species. The fact that arrow was the preferred species was not surprising given that it had the largest leaves and shortest/smallest culms of species offered. This may make arrow bamboo easier to process than the other species. We are currently conducting nutritional analyses to determine if there are nutritional differences between arrow and the other two species.

Conservation Implications

In the wild giant pandas are known to feed on a variety of bamboos (Taylor, 1993) and their preference for different bamboo species changes seasonally (Taylor & Qin, 1987). In conjunction with research on nutritional value of bamboo species, giant panda preferences for different bamboos across seasons should be considered when allocating land for reserves or planning reintroduction programs. Ideally, reserves created in the future will contain an assortment of complimentary and preferred bamboos to maximize giant panda energy intake and ensure species stability. Furthermore, giant pandas should be reintroduced to areas with adequate bamboo species richness and abundance.