, Volume 7, Issue 4, pp 517–525

Use of Infrared Camera to Understand Bats’ Access to Date Palm Sap: Implications for Preventing Nipah Virus Transmission

  • M. Salah Uddin Khan
  • Jahangir Hossain
  • Emily S. Gurley
  • Nazmun Nahar
  • Rebeca Sultana
  • Stephen P. Luby
Original Contribution

DOI: 10.1007/s10393-010-0366-2

Cite this article as:
Salah Uddin Khan, M., Hossain, J., Gurley, E.S. et al. EcoHealth (2010) 7: 517. doi:10.1007/s10393-010-0366-2


Pteropus bats are commonly infected with Nipah virus, but show no signs of illness. Human Nipah outbreaks in Bangladesh coincide with the date palm sap harvesting season. In epidemiologic studies, drinking raw date palm sap is a risk factor for human Nipah infection. We conducted a study to evaluate bats’ access to date palm sap. We mounted infrared cameras that silently captured images upon detection of motion on date palm trees from 5:00 pm to 6:00 am. Additionally, we placed two locally used preventative techniques, bamboo skirts and lime (CaCO3) smeared on date palm trees to assess their effectiveness in preventing bats access to sap. Out of 20 camera-nights of observations, 14 identified 132 visits of bats around the tree, 91 to the shaved surface of the tree where the sap flow originates, 4 at the stream of sap moving toward the collection pot, and no bats at the tap or on the collection pots; the remaining 6 camera-nights recorded no visits. Of the preventative techniques, the bamboo skirt placed for four camera-nights prevented bats access to sap. This study confirmed that bats commonly visited date palm trees and physically contacted the sap collected for human consumption. This is further evidence that date palm sap is an important link between Nipah virus in bats and Nipah virus in humans. Efforts that prevent bat access to the shaved surface and the sap stream of the tree could reduce Nipah spillovers to the human population.


batsdate palm sapfood contaminationinfrared cameranipah virustransmission


Nipah virus is an emerging zoonotic paramixovirus causing encephalitis with 40–70% case fatality in humans (Goh et al., 2000; Luby et al., 2009). Fruit bats of the genus Pteropus, are the apparent natural reservoir for the virus, and occasionally secrete Nipah virus in their saliva and excrete it in their urine (Young et al., 1996; Yob et al., 2001; Olson et al., 2002; Wacharapluesadee et al., 2005). Studies in Bangladesh, India, and Malaysia also identified neutralizing antibodies against Nipah virus in Pteropus bats (P. vampyrus and P. hypomelanus) (Yob et al., 2001; Epstein et al., 2008; Homaira et al., 2010). P. giganteus, one of the most common fruit bats in Bangladesh, have been found to carry antibodies against Nipah virus (Bates and Harrison, 1997; Hsu et al., 2004; Epstein et al., 2006).

Eight Nipah virus outbreaks have been recognized in Bangladesh between 2001 and 2008 (ICDDR,B, 2008; Luby et al., 2009). All of these outbreaks in Bangladesh coincided with the date palm sap harvesting season that occurs from mid-October to early April (Nahar et al., 2010; Luby et al., 2009). In the January 2005 human Nipah outbreak in Tangail District, Nipah cases were eight times more likely than controls to report drinking raw date palm sap (P < 0.001). Date palm sap collectors in the area reported that the large Pteropus fruit bats frequently visited the date palm sap collection pots (Luby et al., 2006).

Drinking raw date palm sap is a seasonal delicacy in Bangladesh. Local sap collectors (gachhis) harvest the sap by shaving approximately 1 inch depth of the bark approximately 8 × 12 inches on one side of the top of the date palm tree (Phoenix sylvestris) just below the base of the branches. The sap collectors shave the bark in a V-shape so that the sap flows into the tap and then into a clay pot that is hung below (Fig. 1.1.a). In Bangladesh, people used date palm sap for two primary purposes: most is made into molasses and a small amount is consumed raw. During winter, the gachhis collect sap from date palm trees early in the morning; they sell it within a few hours to consumers who drink it almost immediately, because it ferments within a few hours and loses its sweet taste (Luby et al., 2006; Nahar et al., 2010).
Figure 1

Date palm sap-producing tree, and ways of bats contaminating the shaved surface of the tree.

Date palm sap collectors identified bats as one of the nuisances that decrease the quality of raw sap. They report that bats drink the sap and, occasionally, spoil the contents of the sap collection pot with urine or feces (Nahar et al., 2010). Traditionally, a few of them use techniques that they thought could prevent bats’ access to date palm sap. A review of these techniques concluded that smearing lime (CaCO3) on the sap stream and covering the sap-producing surfaces with a bamboo skirt, comprised of several bamboo sticks woven together, were the two techniques most likely to prevent bats access to date palm sap (Nahar et al., 2010).

This pilot study aimed to understand the frequency and ways that bats contacted date palm sap, and to gain a preliminary understanding of the effectiveness of these locally used preventive techniques.

Materials and Methods

Site Selection

We conducted the study from February 26 to March 21, 2008 in Faridpur District, Bangladesh, at a location (N23° 33′ 55.0″ E89° 42′ 26.0″) which had about 1000 date palm sap-producing trees and a bat roost (N23° 34′ 19.0″ E89° 42′ 18.0″) containing approximately 200 Pteropus bats, located 1 km away.

Date Palm Tree Selection

Among the date palm sap-producing trees, we selected 14 tall, newly shaved trees that the local sap harvesters’ and residents reported were often visited by bats.


Direct Human Observation

A trained data collector observed a tree each night from dusk (5:00 pm) to dawn (6:00 am). In total, they observed 14 trees over 14 nights, 1 night per tree. The data collector observed the selected date palm trees, maintaining a distance of approximately 30 m from the tree, so as not to alter the bats’ natural feeding behavior. The observer noted the number of bat visits on and around the date palm tree, the number of bats feeding on and around the date palm tree, and the frequency of bats visiting the tree while recording the time of each observation.

Camera Observation

We used two motion sensor-tripped infrared cameras (Silent Image Model RM30 digital cameras, RECONYX, Inc., Holmen, WI)1 that included an infrared illuminator and a passive infrared motion detector. We observed 10 date palm sap-producing trees for 20 camera-nights without any local protective technique. We defined a camera-night of observation as an infrared camera focused on the shaved surface, sap stream, tap, and the collection jar hung on the date palm tree from 5:00 pm in the evening to 6:00 am in the morning. We used two cameras per night, both of which were mounted on a single tree, set to the left and right angle to the shaved surface, and focused on the sap flowing from the shaved part into the collection pot (Fig. 2). If there was any movement in the sap collection area, each camera took one picture per second for the next 5 seconds. The cameras were placed in the same tree that the people were observing, but because there were two cameras, they recorded two camera-nights of observation for each night of human observation. In the morning, these cameras were recovered and the pictures were transferred to a computer and reviewed.
Figure 2

Two infrared cameras mounted in a date palm tree from the left and right angle to the shaved surface, focused on the sap flowing from the shaved surface into the collection pot.

Applied Local Protective Techniques

In addition to the previously described camera observations, we also observed two local techniques, lime paste and bamboo skirts for eight camera-nights. These methods were used to explore if they could prevent bats access to sap. We applied a thin layer of lime paste (CaCO3) over the cut surface on two trees (four camera-nights of observation). We also applied bamboo skirts covering the shaved surface, sap stream, tap, and the opening of the pot on another two trees (four camera-nights of observation) (Fig. 3).
Figure 3

Bamboo skirt, covering a date palm sap-producing tree’s shaved surface, sap stream, tap, and the mouth of the pot.

Data Extraction

We defined bat visits by both human and camera observation. For human observation, this was any sighting of a bat flying in or around the date palm tree. For camera observation, this was when a bat was seen in the focus area of the camera. The visit continued until the bat was no longer seen. We recorded the camera data in a structured field data sheet, which included the frequency of bat visits and any contact with date palm sap. We defined bat–sap contact as an instance of a bat landing, licking (i.e., the bat’s tongue contacted the date palm sap), or urinating either on the date palm’s shaved surface, sap stream, or the parts of the tap or collection pot that came in contact with the sap. The focus of each of the two cameras placed on the same tree did not always fully overlap, so the total bat visits identified by each camera were counted individually. Because each bat could not be individually identified, it was not possible to assess if the same bat made multiple visits at different times of the night. We analyzed the results as the frequency of “bat visits” per “camera-nights” of observation.

We classified the bats into Pteropus and non-Pteropus groups. We distinguished the Pteropus bats by their larger body mass, dog or fox-like face, long and pointed ears, and fur characteristics (Kunz and Jones, 2000a, b), and the non- Pteropus bats based on their relatively smaller body mass and fur characteristics (Nowak et al., 1994). The infrared camera we used only produced low resolution black and white photos, so we were not able to clearly distinguish the species of non-Pteropus fruit bats.

Statistical Methods

We calculated the total, mean, and median bat visits per tree per camera-night of observation. Distribution of the duration of bat–sap contact was skewed, so we presented the medians of the duration of contact with 95% CI (Bonett and Price, 2002). We also performed Wilcoxon rank sum test among Pteropus and non-Pteropus bats’ durations of bat–sap contacts to assess equality of medians (Siegel and Castellan, 1988).


Frequency of Bat Visits to Date Palm Tree

Out of 20 camera-nights of observation of the date palm trees without any intervention, 14 identified 185 bat visits on and around the tree: 151 (82%) at the shaved surface, 4 (2%) at the sap stream, no bats at the tap or the collection pot hung on the tree, and the rest (n = 30) did not land on the tree; the remaining 6 nights of observation did not identify any bat visiting the trees. Out of 20 camera-nights of observations on date palm trees without any intervention, the cameras identified 132 visits of bats at and around the tree (mean: 7.7 visits per night; standard deviation [SD] 15 visits), 91 bat visits at the shaved surface (mean: 7.6 visits per night; SD 8 visits), and 4 bat visits at the sap stream (mean: 0.7 visits per night; SD 1.6 visits).

Human observers during 10 nights without any intervention reported 72 visits of bats (mean: 7.2 visits per night; SD 5.5 visits) in and around the tree. Their observation was unable to identify whether bats landed on the shaved surface, sap stream, tap, or the collection pot and contaminated the sap.

Characteristics of Bats Contacting Date Palm Sap

Pteropus and non-Pteropus bats accessed date palm sap differently. The non-Pteropus bats, primarily landed directly on the shaved surface (n = 82, 54%) (Fig. 1.2). The second most frequent mode of accessing the sap was to land directly to the left or right of the shaved surface (n = 63, 42%) of the tree and climb to it (Fig. 1.3). The Pteropus bats accessed the shaved surface by landing on higher branches and climbing down to the shaved surface (n = 8, 100%) (Fig. 1.4).

Both Pteropus and non-Pteropus bats contacted sap at the date palm trees shaved surface and the sap stream 157 times during a total of 28 camera-nights of observations (mean 5.8 bat–sap contacts per night; SD 11.7). All (n = 8) of the Pteropus bats contacted sap at the shaved surface of the date palm tree by licking (Table 1).
Table 1

Visits, Ways, and Duration of Pteropus and Non-Pteropus Bats Contacting Date Palm Sap

28 Camera-nights of observations




Bats on and round the date palm tree

 Times landed on the tree

8 (4%)

149 (81%)

157 (85%)

 Times did not land on the tree

4 (2%)

24 (13%)

28 (15%)


12 (6%)

173 (94%)

185 (100%)

Frequency of bats contacting date palm sap

 Shaved surface (bat visit-times)

8 (5%)

145 (92%)

153 (97%)

 Sap stream (bat visit-times)

0 (0%)

4 (3%)

4 (3%)


8 (5%)

149 (95%)

157 (100%)

Ways of bat–sap contact

 By landing and licking (bat visit-times)

8 (5%)

130 (83%)

138 (88%)

 By landing only (bat visit-times)

0 (0%)

19 (12%)

19 (12%)


8 (5%)

149 (95%)

157 (100%)

Duration of bat–sap contact

 No. of observations




 Total duration of contamination (seconds)




 Median duration in seconds (range)

483 (124–1094)

20 (1–711)

29 (1–1094)

 95% CI for median




During one camera observation, a Pteropus bat urinated while hanging on the branches of the tree. The urine did not directly contact the shaved surface or the pot, because the bat was hanging on a branch on the right side about 2 feet away (Fig. 4).
Figure 4

A bat urinating while hanging on the date palm tree branch.

Duration of Visits on the Tree and Bats’ Contact with Sap

Bats visited the trees from 1 second to 93 minutes (median: 33.5 seconds). The median duration for the Pteropus bats contacting date palm sap (483 seconds, 95% CI 216–1085) was 23 times higher than the non-Pteropus bats (21 seconds (95% CI: 11–41) (P = 0.01) (Table 1).

Bats Visit at Different Times of Night

Bats visited date palm trees from 6:00 pm to 5:00 am. Camera observation identified that the bats visited the trees most commonly from 9 pm to midnight (37%) and from 2:00 am to 5:00 am (45%) (Fig. 5). The number of Pteropus fruit bat visits (n = 12) were very few compared to the non-Pteropus fruit bat visits (n = 173) (Table 1), and all of the Pteropus fruit bats visited between 8:00 pm and 12:00 am. The cameras consistently identified more bat visits (mean: 5 visits per camera night) than the human observers (mean: 0.4 visits) per night especially between 2:00 am and 5:00 am (95% CI for mean difference: −8.8 to −0.8).
Figure 5

Bats seen by 28 infrared camera and 14 human observations at different hours of the night around the tree.

Local Protective Techniques and Bat Visits

Out of four camera-nights of observations of date palm trees treated with lime paste, the cameras identified 49 bat visits on and around the tree where no sap contact was noted (mean: 13 visits per night; SD 24 visits) and 56 visits at the shaved surface where bats contacted the sap by landing and/or licking (mean: 15 visits per night; SD 27 visits). On several occasions, bats contaminated sap multiple times on a single visit to the tree. During the four camera-nights of observation on trees protected with the bamboo skirt, one bat visited on and around the tree, and no bats contacted the shaved surface, sap stream, tap, or the collection pot.

The human observer collected data on two date palm trees treated with lime smear and two covered with a bamboo skirt. The observer identified 10 bat visits on and around the tree (mean: 5 visits per night; SD 1.4 visits) when lime smear was applied on two trees for two nights. The observer also identified 12 visits of bats on and around the tree (mean: 6 visits per night; SD 5.6 visits) when the trees were protected with bamboo skirt intervention and observed for two nights.


As one step to understanding the transmission of Nipah virus from bats to humans, this study used infrared cameras to observe and explore the frequency and process of bats making physical contact with date palm sap. The infrared photos, taken in areas of previous Nipah outbreaks in Bangladesh, identified that multiple species of bats frequently visited date palm trees and contacted sap by landing on the shaved surface and licking the sap stream.

The majority of the bats physically contacted date palm sap by landing and licking (88%) and, in most cases, it was the shaved surface of the tree. This provided the bats with more surface area to hold on to and gave them easier access to lick the date palm sap (Fig. 1). Physical contact of bats with date palm sap allows microbiological flora from bats to enter sap. Bats that licked the shaved surface potentially contaminated the sap through the transfer of their saliva, though in the vast majority of these instances, we would not expect this flora to include Nipah virus; in capture studies of wild Pteropus bats, fewer than 1% of throat swab specimens had Nipah virus RNA detected (Chua et al., 2002, Wacharapluesadee et al., 2005).

These camera observations also suggest that urine is a potential vehicle of transmission for Nipah virus. Nipah virus RNA and whole virus has been recovered from Pteropus bat urine (Yob et al., 2001; Wacharapluesadee et al., 2005, 2010; Wacharapluesadee and Hemachudha, 2007). Our study results differ from a previous study, which reported that fruit bats do not urinate at their feeding ground (Chua et al., 2002). Our observation of fruit bats, particularly the flying foxes urinating at their feeding ground (Fig. 4), may be due to their rapid gastrointestinal transit time, highly liquefied diet, the long duration of stay, their large consumption of date palm sap (Hall and Richards, 2000), or they could have fed elsewhere prior to their date palm tree visits. The urine might come into contact with the bats’ body fur and feet, and be transferred to sap when the bats land directly on the shaved surface and/or the sap stream of a date palm tree. Another possibility is that bats could urinate on the shaved surface, sap stream, tap of the tree, and/or the collection pot hung on the tree, though we did not directly observe this in 28 camera-nights of observations.

Our camera observations made detailed records of the frequency and timing of bat visits to the date palm tree at night. Previous studies report that P. giganteus bats begin to leave their roost 15 to 20 minutes after sunset and may fly up to 50 km searching for feeding grounds (Lawrence, 1939; Davis, 1966; Marshall, 1985; Kunz and Jones, 2000b). Our findings indicate similar patterns of feeding behavior; the bats started to visit the date palm tree about 1 hour after sunset. This study also identified a distinct feeding pattern, where frequency of bats’ visits peaked twice during the night (Fig. 5), leading to the possibility that they leave the roost several times for foraging, or different bats visited the tree from different distances (Singaravelan and Marimuthu, 2004). This also suggests that interventions for sap protection from bats should be in place for the entire night. The mean number (7.7 visits; SD 15) of bats that visited the date palm sap varied from what was observed in another study of nectar-feeding behavior of several species of fruit bats which found that Cynopterus sphinx mean visits varied from 6.6 (SD 2.6) to 39.3 (SD 14.2) visits every night, and for P. giganteus, it was 2.9 (SD 2.1) to 17.3 (SD 5.1) visits in three locations under natural conditions (Singaravelan and Marimuthu, 2004). Bats forage on date palm sap as they do on nectar and fruit juices (Banack, 1998; Singaravelan and Marimuthu, 2004). Date palm sap is an agricultural product that would not be available for bats in natural settings. This indicates that these bats have become opportunistic feeders during winter in Bangladesh; they might drink date palm sap for food, as well as for rehydration (Walton and Trowbridge, 1983). Continuous deforestation and integrated farming have reduced the availability of naturally occurring food sources including flowers, nectars, pollens, and fruits (Kunz and Jones, 2000a,b; Khan, 2001). Although both nectar and date palm sap are liquid plant- produced sources of food, and easy to access, the high sugar content and the availability of the date palm sap might promote higher frequency of bat visits (Aidoo et al., 2006). The date palm trees usually produce more than a liter of sap/tree/day that consist of about 10% sucrose (FAO, 1993). Whereas, the volume of nectar production is generally much lower (ranging from 0.3 to 103 μL in a day) depending on the species of tree with a sucrose content ranging from 10% to 100% (Wolff, 2006). The date palm sap production starts to decline in mid-February, during the beginning of Spring in the Indian subcontinent, and it is about the same time when nectar becomes available (Mitra et al., 2010).

This study identified that multiple species of fruit bats physically contacted date palm sap, including Pteropus spp. The only Pteropus species present in Bangladesh is P. giganteus (IUCN, 2000). Researchers found that both Pteropus and non-Pteropus bats were serologically positive in several studies in Asia (Yob et al., 2001; Hsu et al., 2004; Li et al., 2008) and Africa (Hayman et al., 2008); in Thailand, Nipah virus RNA was identified in the saliva of Pteropus and non-Pteropus fruit bats (Wacharapluesadee et al., 2005). This indicates that the non-Pteropus fruit bats could also be involved in the transmission of Nipah virus or other zoonoses in Bangladesh. Although the frequencies of non-Pteropus bat visits were higher than the Pteropus fruit bats, the Pteropus bats contacted the sap for a longer time. Further research might clarify the potential role of non-Pteropus bats in zoonotic virus transmission.

To prevent bats from feeding on the date palm sap, the date palm tree harvesters in the western part of Bangladesh apply lime to the shaved surface. The sap collectors who practiced this method believed that the bitter acrid taste of lime would keep bats away (Nahar et al., 2010). In the limited data we collected, the application of lime did not appear to prevent bats from feeding and making contact with date palm sap. The lime might have been washed away by either the sap flow or by rain that occurred during one of the nights of observation. Another preventative technique we observed being used by sap collectors was the bamboo skirts. In the two trees this was tried, the bamboo skirt apparently did protect the sap from bats. The physical barrier formed by the bamboo skirts covering the shaved surface, sap stream, tap, and mouth of the sap collection pot at the tree kept bats from physically contacting the sap.

Bat observation through infrared camera was a useful method to study bat feeding behavior of date palm sap. These cameras provided more accurate detail of bats’ visits than did the human observation. Our human observers did not have optimum night-viewing capabilities to detect bats contacting date palm sap. They may have lost concentration, or fallen asleep in the early hours of the morning, but had a wider field of view. We could only compare bats’ visits on and around the tree between human and camera observations. The human observer noted that the frequency of bats’ visits started to decline after 2 am, and failed to detect any bats after 4:00 am, whereas the camera observation had opposite results. The cameras identified limited numbers of Pteropus bat visits, which may not reflect the actual nature of date palm tree visits at different times of the night. Despite the benefits of the cameras’ exclusive focus on the shaved surface, sap stream, tap, and collection pot, it could have been possible that the cameras took pictures of the same non-Pteropus fruit bat flying around the tree several times due to this narrow view. The non-Pteropus fruit bats have a tendency to forage around fruit trees multiple times (Singaravelan and Marimuthu, 2004), and we identified that most (94%, n = 173) of the bat visits on and around the trees were by non-Pteropus bats. Similarly, the camera might have missed other bats picked up by human observation that were flying around the tree and were out of camera focus. In 4 out of 28 observations, heavy fog reduced visibility for a couple of hours: this could also have reduced our bat count. Our method identifies the frequency of bat visits; this probably underestimates the number of bats visiting each night. Pteropus bats have their preferred feeding grounds where they forage for several hours at night (Walton and Trowbridge, 1983); non-Pteropus fruit bats, such as Cynopterus spp, often make their roost very close to their feeding ground and shuttle back and forth (Storz and Kunz, 1999). This particular feeding behavior of the fruit bats could have led us to count the same bat visiting multiple times.


Bats commonly visited date palm trees and physically contacted sap collected for human consumption. This study provides further evidence that date palm sap is an important link between Nipah in bats and Nipah in humans. Efforts that prevent bat access to the shaved surface and the sap stream of the tree from dusk to dawn could reduce Nipah spillovers to the human population. Further study is needed to identify the efficacy of bamboo skirt to prevent date palm sap contamination by bats.


Inclusion of trade names is for identification only and does not imply endorsement by ICDDR,B, by CDC or the Department of Health and Human Services.



This research study was funded by the United States Centers for Disease Control and Prevention (CDC) grant number 5U51CI00298-04, and the International Centre for Diarrhoeal Disease Research (ICDDR,B) grant number 00357. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the CDC or the ICDDR,B. ICDDR,B acknowledges with gratitude the commitment of CDC to the ICDDR,B’s research efforts. The authors thank our field research officer, Dr. Shahneaz Ali Khan, and date palm sap collector, Mr. Chitta Ranjan, for their efforts in the fieldwork. The authors also thank Dorothy Southern for assistance with writing the manuscript.

Copyright information

© International Association for Ecology and Health 2011

Authors and Affiliations

  • M. Salah Uddin Khan
    • 1
  • Jahangir Hossain
    • 1
  • Emily S. Gurley
    • 1
  • Nazmun Nahar
    • 1
  • Rebeca Sultana
    • 1
  • Stephen P. Luby
    • 1
    • 2
  1. 1.PIDVS, HSID, ICDDR,BDhakaBangladesh
  2. 2.Centers for Disease Control and Prevention (CDC)AtlantaUSA