Ethology Ecology & Evolution 17: 271-277, 2005
An automatic technique for selective feeding and logging of individual wild squirrels B. Kenward
1,3,
R.E. Kenward
2,4
and A. Kacelnik
1,5
1
Department of Zoology, University of Oxford, South Parks Road, Oxford, OX1 3PS, UK 2 Centre for Ecology and Hydrology, Winfrith Technology Centre, Dorchester, DT2 8ZD, UK Received 3 February 2005, accepted 4 July 2005
We describe an automated feeding station for squirrels which identifies individuals and logs their feeding times, using passive integrated transponders (PIT tags). The system can provide or deny food to tagged individuals based on identity and/or arbitrary schedules. We test the system with wild grey squirrels Sciurus carolinensis, and report their feeding patterns and their ability to learn to manipulate a joystick to gain access. Possible applications for the system include behavioural or ecological research where the food access of different free-ranging individuals is to be controlled, and the supplementary feeding of red squirrels Sciurus vulgaris. key words:
feeding patterns, feeding station, grey squirrels, PIT tags, red squirrels.
Introduction
Several studies have demonstrated the use of passive integrated transponders (PIT tags) for identifying wild rodents at automated stations (Harper & Batzli 1996, Dell’Omo et al. 1998). This technology has several advantages when compared with radio-transmitter tags: PIT-tags are much cheaper and no battery is required, so they operate indefinitely and are small enough to implant easily in the field. In a study on over a thousand PIT-tagged ground squirrels, tag loss was less than 5% over the first year and no losses were detected during the second (Schooley et al. 1993). However, unlike radio tags, they cannot be detected at long range, but must be scanned from the range of 10-150 mm. A caught animal can be scanned with a 3 Correspondence: Benjamin Kenward (Tel. +44 1865 271245; Fax +44 1865 310447; E-mail:
[email protected]). 4 Robert Eyres Kenward (Tel. +44 1305 213606; E-mail:
[email protected]). 5 Alex Kacelnik (Tel. +44 1865 271164; E-mail:
[email protected]).
272
B. Kenward, R.E. Kenward and A. Kacelnik
hand-held device, or scanners can be placed in automated stations in which animals pass close to a scanning antenna. A system which, in addition to automated identification of individuals, allowed an automated feeder to provide or deny food access based on identity would have numerous applications. These might include ecological research in which individuals’ food intake could be manipulated, behavioural research in which food rewards were contingent on measurable behavioural responses (with different contingencies for different individuals), and conservation through selective feeding or monitoring of desired individuals. Here we describe such a system, which was developed to study learning in wild grey squirrels Sciurus carolinensis, and we report feeding activity data obtained from a population of grey squirrels over one season.
Description of the system The machine was custom built in the workshops of Oxford University’s Department of Zoology. The system is composed of two modules: the feeding station itself (see Fig. 1) and a control box, connected by cable, which contains a 12 v car battery for power and custom built electronics to control and log the feeding station. The heart of the electronics is a Flashlite microcontroller (JK Microsystems Inc., California, USA) with onboard flash RAM which does not lose data in the event of power failure. The feeding station consists of an entrance tube leading to a feeding chamber, both constructed from transparent acrylate plastic. Looped around the entrance tube is an antenna (Francis Scientific Instruments, Cambridge, UK) which reads the tag of any squirrel passing through the tube. In the feeding chamber is an automated sliding door which can open or close to provide controlled access to a food trough. A plate on the floor toggles a switch when depressed. This allows the detection of the squirrel in the feeding chamber, but also provides a safety mechanism — when the plate is depressed, the door motor circuit is broken so that the door cannot close. A joystick can also be installed in the feeding chamber in front of the door for use as a manipulandum in learning experiments. The system is entirely programmable, so that access to the food can be made contingent on criteria such as identity of squirrel, appropriate manipulation of joystick, time of day, number of times the individual has fed previously, etc. The system is weatherproof and can operate for up to a week before data must be downloaded (onto a portable computer via a serial link) and the battery recharged.
Squirrels enter here
Feeding chamber
Entrance tube Loop antenna
Sliding door Pressure sensing floor
Food trough
Wooden platform Fig. 1. — Side-on schematic diagram of the squirrel feeding station.
Automatic logging of squirrel feeding
273
Monitoring the feeding activity of a grey squirrel population
In 2001 squirrels were trapped in an area of Wolfson College gardens in Oxford with two ground traps, placed approximately 50 m apart and baited with maize. The habitat is wooded parkland with abundant oak, walnut, and other tree species that provide seeds palatable to grey squirrels. During June and July, nine squirrels were tagged during intensive trapping intended to catch and tag all the individuals visiting the area. PIT-tags (size 12 × 2 mm) were injected subcutaneously at the ventral base of the tail, where the tightness of the skin and connecting tissue means they are unlikely to migrate (Kenward et al. 1995). After the trapping period, on July 6th (day 1) the feeding station was installed on a wooden platform next to a tree roughly at the intermediate point between the two traps. An initial habituation period involved spreading hazelnuts liberally on the wooden platform and inside the feeding station to attract squirrels. During this period the hazelnuts were regularly replenished but diminishing amounts were placed outside the machine to encourage the squirrels to enter and travel down the entrance tube to the feeding chamber. The door was left permanently open to allow the squirrels to habituate to the machine without disturbance from moving parts. Of the nine squirrels caught, eight fed at the station during this period, seven of them regularly (Fig. 2). On day 15, the door was programmed to be closed except when a tagged squirrel was identified, at which point it would open to allow access to the feeding chamber. Initially hazelnuts were still placed in the entrance chamber, so that the potentially frightening action of the opening door did not occur before the squirrel had had a chance to get a nut, but the supply of these nuts was gradually diminished. From day 18, hazelnuts were only placed behind the door in the food reservoir area, so that the squirrels could only obtain food by entering the station, causing the door to open in front of them, and going through the door to obtain hazelnuts. The station was left operating in this manner, with the hazelnuts regularly refilled, until day 41. Over this period, five of the eight squirrels which fed initially were regular feeders at the station (Fig. 2). On day 41 a pilot learning experiment was begun, in which the joystick was placed in the feeding chamber to see if the squirrels could learn to use it. From this moment, the door would not open to allow access to the food until a tag had been read, followed by a push on the joystick within 30 sec. During this period, which lasted for 21 days, only one squirrel (squirrel 1 — an adult female) learnt to operate the joystick. Three other squirrels occasionally visited but did not push the joystick and so never fed. Towards the end of this period even the squirrel which had learnt to use the joystick stopped coming (Fig. 2). At various times during the study, activity around the feeding station was recorded by a video camera (XCam). The camera was attached to a tree branch, powered by the station’s 12 v battery, and transmitted wirelessly to a tape recorder in a nearby building. The footage confirmed that the squirrels were operating the machine as intended. It also revealed that social interactions around the feeder were common. Dominant individuals would sometimes defend the feeder aggressively. On other occasions individuals would simultaneously occupy space around the feeder and take turns at feeding without aggression. Figs 3 and 4 show patterns of feeding at the station.
60
40
20
0
1
6
nently
Door perma
11
open 16
Day
21 26
Door o pens o nly when ta g dete cted
31
36
Door o pens w hen ta detecte g d and joystic k move d 41 46 51 56
1 adult female
2 adult male
3 adult male
4 adult male
Fig. 2. — Number of daily visits to the station of the eight squirrels during the 56 days of the whole trial. Day 1 is July 6th. Each phase of the trial is marked. In the final phase, when no food was available unless the joystick was moved, only squirrel 1 fed when it visited because the others did not learn to operate the joystick. The gaps represent times when data was lost due to program error or times when the station was removed for improvement or repair. Alternation between black and white data series is for clarity only.
Number of visits
80
6 adult female
7 juvenile male
8 juvenile female
5 adult male
Individual
274 B. Kenward, R.E. Kenward and A. Kacelnik
Automatic logging of squirrel feeding
275
2.5 1 adult female 2 adult male 3 adult male 4 adult male 5 adult male 6 adult female 7 juvenile male 8 juvenile female
Average number of visits
2
1.5
1
0.5
0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Hour of day Fig. 3. — Average number of visits of eight squirrels to the feeding station by hour of day over the first 41 days of the trial.
8 juvenile female 7 juvenile male
Individual
6 adult female 5 adult male 4 adult male 3 adult male 05:00:00
06:00:00
07:00:00
08:00:00
09:00:00
10:00:00
Time Fig. 4. — Pattern of feeding activity at the station on the morning of day 9, July 14th, chosen as a representative busy day. Each diamond is one visit. The two squirrels not shown did not visit on this day.
Discussion
The system was able to provide food and log the feeding activity of the visiting squirrels. Gaps in service were due to trialling alternative methods of dispensing
276
B. Kenward, R.E. Kenward and A. Kacelnik
food, and to the implementation of minor improvements to the machinery and controlling program because the system was being developed during testing — by the end of the season the system had been developed to a stage where it could operate indefinitely. Eight out of nine squirrels trapped in the area used the system, showing that most squirrels willing to enter a trap were also willing to enter the feeding station. The system is therefore suitable for use, although more trials, and some improvements (see below), may be desirable to ensure reliability. That most squirrels stopped visiting the machine in early September is probably because they failed to learn to operate the joystick and so were no longer being fed. However, this is also the time of year when tree seeds become abundant as alternative food and grey squirrels become hard to catch in ground traps (Dubock 1979), which may explain the cessation of visits by the one squirrel that learned to use the joystick. The daily feeding pattern data shows activity similar to that found by Kenward (1982), with grey squirrels primarily feeding in the early morning and late afternoon. The use of the joystick demonstrated the system’s ability to provide food only on the satisfaction of certain criteria. The failure of most squirrels to learn to use the joystick suggests that this method may not be suitable for learning experiments in wild squirrels. Other future improvements might include such additions as an automatic balance which could weigh each individual (Boisvert & Sherry 2000). One aspect of the feeding station which may have caused some squirrels to avoid it was the noise and movement of the door. There was an overall reduction in number of visits when the door was introduced, and squirrels could be seen on the video recordings showing signs of fear when the door opened. An improvement from this perspective would be to use a hinged door, unlocked by a solenoid only for permitted animals, who could then push it open — standard squirrel traps use push-doors so it is known that squirrels easily learn to operate them. Alternative methods of dispensing food were tested, including a custom built hazelnut dispenser and a laboratory rat-pellet dispenser, but neither was able to function adequately for long periods in the wet outdoor environment. There are a number of possible further applications of this system. It could be ideal for answering ecological questions, for example to tease apart the causal relationships between food availability, range size, and population density (Kenward 1985). Other possible applications relate to the conservation of the Eurasian red squirrel Sciurus vulgaris, in decline in Great Britain, Ireland, and Italy (Bertolino & Genovesi 2003, Gurnell et al. 2004) due to grey squirrel introduction. It is likely that disease transmission between the non-native greys and the reds is at least partially responsible for the decline (Rushton et al. 2000; Tompkins et al. 2002, 2003). If cross infection occurs because of inter-species social interactions, it is important to know whether feeding stations may increase the rate of cross infection by concentrating activity in their vicinity. A feeding station as described here, in combination with video recordings, could offer an automated method to investigate feeding patterns and associated social interactions. Some red squirrel conservation programs employ supplementary feeding, using a food hopper with a see-saw entrance platform which allows access by red squirrels but tips out grey squirrels, which are heavier (Gurnell & Pepper 1993). The system described here would provide an alternative method of selectively feeding tagged red squirrels without the scope for evasion in a weight-based system (e.g. by juvenile grey squirrels).
Automatic logging of squirrel feeding
277
Acknowledgements Thanks to Tony Price and Robert O’Dell of Oxford University’s Zoology Department Engineering Workshop for building the system, Mike Francis of Francis Scientific Instruments for providing PIT electronics and advice, Wolfson College for allowing access to their squirrels, the Dumbleton Trust which funded Benjamin Kenward during this work, the organisers of the Third International Tree Squirrel Colloquium, attendance at which inspired the publication of this work, and one anonymous referee, Luc Wauters, and Louise Magris for comments on the manuscript.
References Bertolino S. & Genovesi P. 2003. Spread and attempted eradication of the grey squirrel (Sciurus carolinensis) in Italy, and consequences for the red squirrel (Sciurus vulgaris) in Eurasia. Biological Conservation 109: 351-358. Boisvert M.J. & Sherry D.F. 2000. A system for the automated recording of feeding behavior and body weight. Physiology & Behavior 71: 147-151. Dell’omo G., Shore R.F. & Lipp H.P. 1998. An automated system, based on microchips, for monitoring individual activity in wild small mammals. Journal of Experimental Zoology 280: 97-99. Dubock A.C. 1979. Methods of age determination in grey squirrels (Sciurus carolinensis) in Britain. Journal of Zoology, London (A) 187: 27-40. Gurnell J. & Pepper H. 1993. A critical look at conserving the British red squirrel Sciurus vulgaris. Mammal Review 23: 127-137. Gurnell J., Wauters L.A., Lurz P.W.W. & Tosi G. 2004. Alien species and interspecific competition: effects of introduced eastern grey squirrels on red squirrel population dynamics. Journal of Animal Ecology 73: 26-35. Harper S.J. & Batzli G.O. 1996. Monitoring use of runways by voles with passive integrated transponders. Journal of Mammalogy 77: 364-369. Kenward R.E. 1982. Techniques for monitoring the behaviour of grey squirrels by radio, pp. 175-196. In: Cheeseman C.L. & Mitson R.B., Edits. Telemetric studies of vertebrates. Symposia of the Zoological Society of London No. 49. Kenward R.E. 1985. Ranging behaviour and population dynamics in grey squirrels, pp. 319330. In: Sibley R.M. & Smith R.H., Edits. Behavioural ecology: ecological consequences of adaptive behaviour. Oxford: Blackwell Scientific. Kenward R.E., Hodder K.H. & Rose R.J. 1995. Conservation of red squirrels: second annual report. Furzebrook, UK: Institute of Terrestrial Ecology. Rushton S.P., Lurz P.W.W., Gurnell J. & Fuller R. 2000. Modelling the spatial dynamics of parapoxvirus disease in red and grey squirrels: a possible cause of the decline in the red squirrel in the UK? Journal of Applied Ecology 37: 997-1012. Schooley R.L., Vanhorne B. & Burnham K.P. 1993. Passive integrated transponders for marking free-ranging Townsend ground-squirrels. Journal of Mammalogy 74: 480-484. Tompkins D.M., Sainsbury A.W., Nettleton P., Buxton D. & Gurnell J. 2002. Parapoxvirus causes a deleterious disease in red squirrels associated with UK population declines. Proceedings of the Royal Society of London (Series B, Biological Sciences) 269: 529-533. Tompkins D.M., White A.R. & Boots M. 2003. Ecological replacement of native red squirrels by invasive greys driven by disease. Ecology Letters 6: 189-196.
