OUP user menu

Dietary Characteristics of Myotis ricketti in Beijing, North China

Jie Ma, Jinshuo Zhang, Bing Liang, Libiao Zhang, Shuyi Zhang, Walter Metzner
DOI: http://dx.doi.org/10.1644/05-MAMM-A-183R1.1 339-344 First published online: 21 April 2006


We evaluated composition and seasonal variation in the diet of Rickett's big-footed myotis (Myotis ricketti) by examining 342 fecal samples collected every 2 weeks when the bats were active in Fangshan District of Beijing, North China, from 2002 to 2003. The diet consisted of 3 kinds of fish (Zacco platypus, Carassius auratus, and Phoxinus lagowskii) and at least 7 orders of insects (Coleoptera, Lepidoptera, Homoptera, Ephemeroptera, Hemiptera, Diptera, and Hymenoptera). Fragments of fish accounted for 67.1% and fragments of insects for 28% (by volume) of the diet; Z. platypus was the dominant food (60.1%), followed by Coleoptera (13.1%). Significant differences existed between the proportion of fish and insects in the diet; however, no seasonal variation in the diet was found over the course of the investigation. This suggests that at this study site in Beijing, M. ricketti was highly specialized in foraging on only 1 species of fish and relied less on insects as alternative food sources, despite their seasonally high abundance.

Key words
  • Beijing
  • China
  • diet
  • fecal analysis
  • Rickett's big-footed bat

Empirical data about fecal composition have been widely used to gain insights into the diet and foraging habits of insectivorous bats (Kurta and Whitaker 1998). Some findings that are based on comparisons of fecal composition with insect abundance in the feeding area of particular bat species suggest that insectivorous bats feed selectively on certain groups of insects (Agosta and Morton 2003; Anthony and Kunz 1977; Hurst and Lacki 1997; Schulz 2000; Swift et al. 1985; Vaughan 1997). In contrast, others question the reliability of this conclusion (Whitaker et al. 1999) and argue that insectivorous bats feed opportunistically (Belwood and Fenton 1976; Fenton 1982; Fenton and Morris 1976; O'Neill and Taylor 1989) or specialize on particular insect groups only at certain times of the year (Anthony and Kunz 1977; Fenton and Thomas 1980). Therefore, Whitaker et al. (1999) suggested that the sampling methods used for examining the composition of the diet and the availability of food must be carefully considered before classifying insectivorous bats as specialists or opportunists.

Compared with data in insectivorous bats, dietary composition and its potential seasonal variation in piscivorous bats are still unknown. To date, 3 bat species show fishing behavior in the field, Noctilio leporinus (Bloedel 1955; Novick and Dale 1971; Schnitzler et al. 1994), Myotis vivesi (Altenbach 1989; Reeder and Nortis 1954), and Myotis ricketti (Rickett's big-footed myotis—Ma et al. 2003). Fishing behavior also has been observed for a few species in the laboratory, such as in Noctilio albiventris (Suthers and Fattu 1973) and Myotis daubentonii (Siemers et al. 2001a). Brooke (1994) and Blood and Clark (1998) examined the diet of N. leporinus and M. vivesi, respectively. They found that fish occurred in 53% (by volume) of the diet of N. leporinus on the basis of a 9-month investigation (Brooke 1994). In contrast, M. vivesi appeared to have consumed more crustaceans than fish (Blood and Clark 1998). The 3rd piscivorous bat species, M. ricketti, had long been suspected to feed on fish, mainly based upon morphological features, such as its very large feet (Allen 1936; Thomas 1894). Some studies have attempted to determine the amount of fish eaten by this bat by analyzing fecal pellets and found it to be approximately 50% (by volume) in the diet (in Laos by Robinson and Webber [1998] and in Hong Kong by Ades [1999]). However, these studies in M. ricketti were based upon a very limited sample sizes. Furthermore, all studies in piscivorous bats, including our own previous study (Ma et al. 2003), so far represent only snapshots of the bats' activity exhibited during the entire season and furthermore were taken from very different locations covering a wide range of habitat types, from tropical (Laos) and subtropical (Hong Kong) broad-leafed deciduous forests to warm, temperate, mixed forests (Beijing). Therefore, we aimed at obtaining a long-term picture of the dietary habits of a colony of M. ricketti by assessing food composition and its seasonal variation in a known foraging area (Ma et al. 2003) over a period of 2 years.

Materials and Methods

Study site.—Fecal samples and reference specimens were collected from a cave in the Taihang Mountains near Sihe Village (altitude: 900-1,200 m) in Fangshan County, District of Beijing, China (Fig. 1). The cave, known locally as the Bat Cave (115°59′N, 39°43′E), where M. ricketti was roosting also was occupied by 3 other bat species: large mouse-eared bat (Myotis chine nsis), greater horseshoe bat (Rhinolophus ferrumequinum), and greater tube-nosed bat (Murina leucogaster). M. ricketti was the dominant species, with 1,000-2,000 individuals. The other species were represented by about 100 individuals each. This cave was the only site in an area of several hundred square kilometers where we found M. ricketti roosting. The mean yearly temperature in the study area was 7–15°C and the annual rainfall was 500–600 mm, concentrated in the wet season from mid-June to the end of August (wet season); all other months have little rainfall (dry season—Mao and Song 1997). For the purpose of comparing food item distributions in different seasons, spring was defined as from March to May, summer from June to August, and autumn from September to November.

Fig. 1

Map indicating location of study site in Fangshan County in southwestern portion of the District of Beijing, North China. Bat Cave is site of sampling.

A reservoir (Fig. 2) fed by the Liuli River and covering a surface area of at least 30,000 m2 was located approximately 8 km southeast from the cave and contained at least 10 fish species (including Zacco platypus, Carassius auratus, Phoxinus lagowskii, Cyprinus carpio, Silurus asotus, and Aristichthys nobilis). The areas surrounding the cave and the reservoir are covered by warm, temperate, mixed forest, which included Chinese pine (Pinus tabulaeformis), arborvitae (Sadina chinensis), oak (Quercus mongolica and Quercus liaotun-gensis), and aspen (Populus davidiana).

Fig. 2

Reservoir and adjacent habitat structure, near study site, showing A) the dam and B) the upriver end of the reservoir. Location of reservoir is shown in Fig. 1.

Collecting fecal samples and classifying food categories.—Myotis ricketti was active from mid-March to early November and hibernated during the rest of the year in the Bat Cave. Fecal samples were collected every 2 weeks from bats collected in mist nets at the cave entrance from May to November 2002 and April to November 2003. (We did not attempt to collect fecal samples during hibernation in order not to disturb the bats.) Bats were captured when they had finished foraging and returned to the roost and were kept individually in cloth bags for a maximum of 5 h and brought to Sihe village, because it was not practical to remain near the cave entrance until all fecal pellets were collected. Care was taken to keep the bags with the bats in a warm place with relatively high humidity, especially in autumn. Before we released the bat, all droppings from 1 bag (i.e., from 1 bat) were combined and treated as 1 sample. Bats were released at the site of capture early the following morning. The bags were subsequently washed to ensure they were clean for the next collection of fecal samples 2 weeks later. Bats were handled and released according to guidelines established by the American Society of Mammalogists (Animal Care and use Committee 1998) and procedures were approved by the Animal Research Committee at the University of California, Los Angeles.

In every collection, all fecal pellets of 1 bat were collected as 1 sample and stored in 70% ethanol in small plastic vials. Collections of reference insects and fish were obtained in the field and were used to aid in identification of fecal contents. Each sample collected was examined separately under a stereomicroscope in the laboratory and the percentage volume of each prey category in each sample was estimated visually to the nearest 5%. Insect remains were identified to the level of order (Zheng and Gui 1999) and fish to species level (Wang 1984). Data for each collection were then presented as overall percentage volume (sum of individual volumes for each food in each sample/total volume for the samples × 100), indicating the total amount of each food in the samples (Whitaker et al. 1999; Whitaker and Rodríguez-Durán 1999). Percentage volumes indicate the relative amount of each food in each sample (Whitaker 1988).

In addition, we also recorded the echolocation signals of bats in the study area using a bat detector (D980, Pettersson Elektronic AB, Uppsala, Sweden).

Data analysis.—In our study, all percentage volumes were transformed into arcsine values to achieve normality for statistical analysis using a commercial statistics analysis program (SPSS 10.0— Whitaker et al. 1999; Zar 1984). A 2-tailed paired t-test was performed to compare the difference between fish and insect volumes and between the same seasons in the 2 years. One-way analysis of variance (ANOVA) with Student-Newman-Keuls tests was then applied to the data to test for differences in the foods (fish and insects) collected at different dates. The diet diversity was calculated between wet and dry season using Shannon's index of diversity, H = ΣPi InPi, where Pi is the proportion of the ith prey category out of all prey species; the difference between wet season and dry season was analyzed with a chi-square test (Sinclair and Zeppelin 2002). In addition, a chi-square test also was used to test the differences in the average number of prey taxa (i.e., sum of number of species of fish and number of orders of insects divided by collection times) among different seasons.


Principal diet.—The diet of M. ricketti consisted of 3 species of fish and 7 orders of insects based on a total of 342 fecal samples over the 2 years (Table 1). Throughout both years, the pale chub (Z. platypus, Cypriniformes, Pisces) was the most common food item in the bats' diet. It comprised more than 50% in 21 of the 28 samples (taken every 2 weeks), with 85 samples (24.9% of all samples) consisting entirely of this fish species. In summary, pale chubs occurred in 60.2% of fecal samples (206 of 342), composing 60.1% (by volume) of diet. The consumption of pale chubs decreased gradually over the course of each season, reaching the lowest point in early August in 2002 and mid-June in 2003, but increased again in the subsequent months in each year. The other 2 species of fish found in the diet of M. ricketti, although at much lower percentage volumes, were goldfish (C. auratus) at 5.0% and Amur minnow (P. lagowskii) at 2.0%.

View this table:
Table 1

Volume percentages of fecal composition of Myotis ricketti from Beijing, North China, during the observation period.

YearCollectionnZacco platypusCarassius auratusPhoxinus lagowskiiColeopteraLepidopteraHomopteraEphemeropteraHemipteraDipteraHymenopteraUnidentified insectsTotal
20021 May875.000000015.06.503.099.5
14 May1266.70012.3000017.001.097.0
2 June950.00026.85.010.000002.093.8
17 June1050.010.0023.0000010.206.299.4
1 July1233.38.3030.
16 July1030.010.0022.05.018.0010.
3 August1225.008.335.015.510.005.0001.099.8
16 August1346.215.4020.05.013.00000099.6
31 August670.00012.
14 September657.
28 September660.08.0015.007.208.0001.099.2
13 October2575.0012.55.0004.003.00099.5
1 November121000000000000100
20031 April1283.38.30006.000000.498.0
15 April885.08.803.001.00000097.8
1 May1060.010.0018.005.00003.01.097.0
16 May2170.04.8005.510.81.03.0000.996.0
3 June2155.04.8010.
17 June1225.008.325.08.515.01.312.0002.797.8
2 July1441.77.1025.112.84.8003.101.496.0
16 July1545.00020.
1 August1585.0009.006.000000100
15 August1258.3010.08.75.8004.505.03.796.0
2 September1550.010.0020.
16 September1160.010.0015.06.03.000003.097.0
2 October1170.
16 October1075.01007.05.0000001.098.0
1 November1490.006.03.00000001.0100

The 2nd most prominent food item was beetles (Coleoptera), with a total of 13.1% (volume) for both years. On 10 of 28 collection dates, in summer from June to August, beetles accounted for more than 20% of percentage volume (Table 1). Other insect orders occurred less often in the bats' diet. In decreasing order of percentage volume, they were leafhoppers (Homoptera, 5.6%), moths (Lepidoptera, 4.3%), true bugs (Hemiptera, 2.6%), flies (Diptera, 1.0%), flying ants (Hymenoptera, 0.9%), mayflies (Ephemeroptera, 0.5%), and unidentified insect remains (1.5%). Hence, with the exception of pale chubs (Z. platypus) and beetles (Coleoptera), most food items represented less than 10% in the diet of M. ricketti Table 1).

The total percentage volumes of fish and insects for the 2 years were 67.1% and 28.0%, respectively (Table 2). The proportion of fish in the diet was significantly higher than that of insects in each year. However, no difference was found between the same food groups (fish to fish and insect to insect) in each year (paired t-test, P > 0.05 for both comparisons; Table 2). Volume in percentages for fish might somewhat underestimate the actual proportion of fish consumed by the bats because vertebrates contain more digestible material (e.g., muscles) than do insects. Therefore, our data represent a conservative estimate of the differences between fish and insects. However, this does not affect the trend in the data presented here.

View this table:
Table 2

Comparisons between the total amounts of fish and insects in the diet of Myotis ricketti, shown as percentage volume. All differences are significant at P < 0.01 (t-test).

YearnVolume % (X̄ ± SD)Volume % (X̄ ± SD)t
20021363.5 ± 19.033.8 ± 18.32.607
20031571.2 ± 19.024.1 ± 17.04.825
2002 and 20032867.1 ± 19.028.0 ± 18.05.202

Seasonal variation.—This clear overall preference of M. ricketti for fish as the main diet item in our study area remained the same throughout the seasons in both years; no significant seasonal variation was found in either year (1-way ANOVA, 2002: F = 0.206, d.f. = 2, P = 0.815; 2003: F = 1.302, d.f. = 2, P = 0.289). Also, no differences were found between the same seasons for the 2 years (spring, summer, and autumn; paired Mest, 3 tests, all P > 0.05). The proportion of fish in each year was higher in the dry seasons than in the wet seasons. Fragments of fish accounted for ≥50% (by volume) in the dry seasons in both years, but they dropped below 50% in the wet seasons with the exception of August 2003. However, statistical analysis showed that the difference was not significant when both years were combined (1-way ANOVA, F = 0.023, d.f. = 3, P = 0.995). Conversely, insect components increased rapidly in the wet season of each year. Nevertheless, overall no difference was found in biodiversity between wet and dry seasons (chi-square test, 2002: χ2 = 0.01, P > 0.05; 2003: χ2 = 0.46, P > 0.05).

However, we did find some differences between seasons in species of fish and insect orders. Although pale chubs (Z. platypus) and beetles (Coleoptera) remained the major food items among fish and insects, respectively, amid fish, goldfish (C. auratus) and Amur minnows (P. lagowskii) were only identified in summer and autumn. Among insects, moths (Lepidoptera) were most abundant in the diet in summer and early autumn, true bugs (Hemiptera) mostly in summer, leafhoppers (Homopterans) mainly from June through September, and Diptera, Ephemeroptera, and Hymenoptera were commonly found in the diet until August of each year.

Finally, when estimating the average number of prey taxa, that is, the combined number of species of fish and number of orders of insects divided by the number of collection times for a particular season, we also found some differences (chi-square test, 5 tests, all P > 0.05). In both years, the bats seemed to feed on a larger number of different food items in summer, whereas the numbers were lower in autumn and lowest in spring. The average number of prey taxa were 5.3 and 5.8 in summer of 2002 and 2003, respectively, which were higher than those in spring (3.0 in 2002 and 4.4 in 2003) and autumn (4.3 in 2002 and 4.8 in 2003). Overall, the average number of prey taxa in dry seasons (4.0 in 2002 and 4.9 in 2003) was lower than that in both wet seasons (5.6 in 2002 and 5.4 in 2003).


We evaluated the composition and seasonal variation in the diet of M. ricketti over a period of 2 years at a location in Fangshan District of Beijing, North China. Throughout both years, the diet consisted mainly of 3 species of fish, with Z. platypus by far the most commonly caught species, and C. auratus and P. lagowskii representing the other 2 prey species (Table 1). The remaining volume (less than one-third of the total) of the diet was composed of 7 orders of insects, mostly of beetles. Other insect groups contributing to the bat's diet were, in descending order of percentage volume, Homoptera, Lepidoptera, Hemiptera, Diptera, Hymenoptera, and Ephemeroptera. Over the seasons, fish remained by far the main food item although the proportion of insects increased slightly during the wet season when they became most abundant. When pooling all fish species and insect orders into 2 food categories, the average prey taxa and biodiversity index of these food categories were higher in wet seasons (i.e., summers) than those in other seasons. Overall, in our investigation, these differences were not significant, thus indicating a stable dietary composition. Although the data from previous studies (in Laos by Robinson and Webber [1998] and in Hong Kong by Ades [1999]), including our own (Ma et al. 2003), only represented a snapshot in the dietary habits of this bat species, examination of our present data now allows us to firmly establish that at least at our study site, M. ricketti was indeed highly specialized in foraging on fish, which is consistent with its specialized morphological characteristics. Like other trawling bats, M. ricketti not only possesses very large hind feet, ideal for grabbing and holding onto objects picked up from the water surface, it also shares important aspects of its flight morphology with other piscivorous bats. The flight morphology is expressed in terms of wing load (body mass/wing area) and aspect ratio (wingspan2/wing area). Aldridge and Brigham (1988) and Altringham (1996) have pointed out that these 2 variables are essential and determine the foraging behavior of bats by affecting their flight speed, efficiency, and maneuverability. One of us (JM) has calculated that, similar to N. leporinus and M. vivesi (Altringham 1996; Norberg and Rayner 1987), M. ricketti has a low wing load (0.07 g/cm2) and high aspect ratio (7.5). The low wing load allows for a low flight speed and sufficient time to scan for prey close to the water surface. It also enables the bat to take off relatively easily even when carrying heavy prey in its hind feet. Piscivore foraging behavior also is supported by the high aspect ratio, which is advantageous for reducing energy expenditures during flight and long periods of foraging over water.

But why were the bats selecting only 3 species of fish (with Z. platypus being far more numerous than the other 2 species) out of the more than 10 species occurring in the reservoir? A possible explanation for the seemingly selective predation on Z. platypus may lie partly in its year-round abundance in the reservoir: Z. platypus outnumbered all other fish species. In addition, their small body size of 3–6 cm (body mass, 1.5-5 g) and their pelagic lifestyle (Wang 1984) make them a vulnerable prey for the bats, and one of us (JM) has observed that these fish group together in small shoals and swim close to the water surface, especially after sunset, which would also make them vulnerable as prey to bats. In contrast to Z. platypus, the other 2 fish species, C. auratus and P. lagowskii, which were less commonly found in the diet of M. ricketti in our study, are normally benthic and live close to the lake or river bottom (Wang 1984) and thus are less easily detected by foraging piscivorous bats.

Foraging exclusively on fish also is consistent with the echolocation behavior of M. ricketti (Ma et al. 2003), as can be seen by the way they respond to Z. platypus. As a result of the pelagic lifestyle of Z. platypus and the fact that they formed small shoals of fish, these fish produce numerous, conspicuous ripples on the water surface with their back fins while swimming and with their mouths when gleaning insects from the surface. The ripples and any exposed body parts of the fish represented prominent echo-reflecting objects on the acoustically smooth water surface (Siemers et al. 2001a; Suthers, 1965, 1967; Wenstrup and Suthers, 1984) and could fit well into the search image of a piscivorous bat. M. ricketti (Ma et al. 2003) and other trawling bats that feed over water all use brief broadband frequency-modulated signals (e.g., M. vivesi [Suthers 1967], N. leporinus [Wenstrup and Suthers 1984], M. daubentonii [Jones and Rayner 1988; Siemers et al. 2001b], and My Otis adver sus [Jones and Rayner 1991]). But it appears that M. ricketti does not rely only on active echolocation to detect fish. Monitoring the bats' echolocation calls with a bat detector revealed that they consistently fell silent immediately before their hind feet touched the water surface until they rose off the water surface back up into the air. Only then did they start emitting search calls again. This period of silence could have been used to listen to noise coming from splashes produced by the fish's movements. It is known that N. leporinus is attracted to sounds produced by splashing water as a result of jumping fish or pebbles hitting the water both in the laboratory and in the wild (Wenstrup and Suthers 1984).

In summary, although examination of our present data indicates that M. ricketti was highly selective in foraging on fish, it is necessary to be careful in generalizing these results. This highly stable diet might merely reflect local conditions at our study site and it is plausible that at different localities where, for example, fish are more intermittently available or insect abundance is greater, these bats exhibit a greater preference toward insects as their principal diet. It is well known that the temporal and spatial distribution of prey in a given area can affect the composition of the diet of bats (Jones and Rydell 2003). For example, it has been shown that the variety and availability of fish determine the proportion of fish in the diet of N. leporinus (Bordignon and Franca 2002; Brooke 1994; Romero 1985). Similarly, the diet of colonies of Eptesicus fuscus in Indiana and Illinois consisted of very few items, but it appeared that this seemingly specialized foraging behavior was merely a result of these bats feeding primarily on insects associated with local crops (Whitaker 1995). Therefore, and despite its morphological and behavioral adaptations, it might be too early or arbitrary to unequivocally assume that M. ricketti is a true fishing specialist and always forages chiefly on fish year-round and in every type of habitat.


This study was funded by grants from the National Institutes of Health (DC005400) and the Whitaker Foundation to WM, the National Natural Science Foundation of China (grants 30025007, 30270169, and 30370264) to SZ, and Innovative Program KSCX2-SW-118 from the Chinese Academy of Sciences to SYZ. We are grateful for help from Z. Tang, J. Zhang, W. Li, T. Wu, J. Chen, Y. Wang, and G. Jia (all Institute of Zoology, Chinese Academy of Sciences, Beijing).


  • Associate Editor was William L. Gannon.

Literature Cited

View Abstract