FORAGING BEHAVIOUR OF ARMADILLIDIUM VULGARE (ISOPODA: ONISCIDEA) IN HETEROGENEOUS
JOANNE M. TUCK and MARK HASSALL1,2)
(Centre for Ecology, Evolution and Conservation, School of Environmental Sciences, University of East Anglia, Norwich, NR4 7TJ, UK)
Foraging behaviour of Armadillidiumvulgare was observed in laboratory arenas in which the spatial distribution of patches of high quality food (powdered dicotyledonous leaf litter) was varied within a background of low quality food (powdered grass leaf litter). The hypotheses that the foraging behaviour and foraging path of A. vulgare would be in uenced by food quality and the patchiness of high quality food resources were tested.
More time was spent in high quality food patches than in low quality food backgrounds than expected by chance in all heterogeneitytreatments, but an increasinglyhigher percentage of time was spent in low quality food as the high quality food became more clumped in space. More time was spent searching, but less time was spent feeding in low quality food backgrounds than in high quality food patches in all the treatments.Walking speed was found to be lower in high quality food patches than in low quality food backgroundsand this was not affected by treatment. Turning frequency and turning angle were found to be higher in high quality food patches than in low quality backgrounds. Turning frequency in low quality food backgrounds decreased as the high quality food became more clumped in space, whereas turning angle in high quality food patches signiŽ cantly increased in the patchy, but then decreased again in the clumped treatment.
The effects of varying the spatial heterogeneity of high quality foods on the trade-off between costs of searching and intake beneŽ ts for saprophages are discussed in relation to predictions from optimal foraging theory for circumstances when intake rate maximisation is affected by the constraint of limited nutrients.
Keywords: optimal foraging, resource patchiness, food quality, terrestrial isopods, woodlice.
1) Corresponding author’s e-mail address: email@example.com 2) This work was funded by a studentship from the Natural Environment Research Council.
© Koninklijke Brill NV, Leiden, 2004 Behaviour 141, 233-244 Also available online –
234 TUCK & HASSALL
The Ž tness of an animal is affected by both the quality and quantity of its diet. For some animals maximizing the energetic gain from foraging is often constrained by the requirements for one or more essential nutrients (Stephens & Krebs, 1986; Bjorndal, 1989), most commonly nitrogen (White, 1993). There may be a trade-off between searching for foods high in nitrogen content and feeding on foods containing the most readily digested energy (Crawley, 1983).
Optimal foraging theory predicts that individuals will evolve foraging strategies that maximize net energy gain, subject to constraints of nutri- ent limitation, through trade-offs between the costs of searching for more proŽ table patches and beneŽ ts of immediate feeding (MacArthur & Pianka, 1966; Emlen, 1966; Krebs et al., 1983). Both theoretical and empirical stud- ies have shown that the behaviour of many predators and herbivores fol- low optimal foraging predictions (Hughes, 1993; Wanink & Zwarts, 1996; Seed & Hughes, 1997; Goldberg et al., 1999). The trade-off between costs of searching and intake gains is not only strongly in uenced by the spatial distribution of food resources (Pyke, 1984; Hassall, 1996; Wallis De Vries et al., 1999) but also for herbivores by the quality of their food (Crawley, 1983).
Food quality is now widely recognized to affect foraging decisions of saprophages (Cadish & Giller, 1997) but almost nothing is known about how decomposers are affected by spatial heterogeneity in quality of nutri- tional resources. Grassland isopods are constrained by the nitrogen content of food (Rushton & Hassall, 1983; Hassall & Rushton, 1984). The terrestrial isopod Armadillidium vulgare both grows and survives better on a diet of dead dicotyledonous leaves than when fed only monocotyledonous material (Rushton & Hassall, 1983), so as both growth and survivorship are closely correlated with Ž tness, dead dicotyledonous food is by deŽ nition a higher quality food than dead monocotyledonous food (Crawley, 1983).
To investigate the effect of spatial heterogeneity in food quality on A. vul- gare the difference in quality of the experimental foods needs to be con- Ž rmed and was studied by monitoring growth rates of A. vulgare on diets of two foods, differing in nitrogen content. The high quality food was naturally senesced leaves from the umbellifer alexanders Smyrnium olusatrum L. and the low quality food consisted of dead leaves from the grass Festuca ovina.
FORAGING BEHAVIOUR IN HETEROGENEOUS ENVIRONMENTS 235
For A. vulgare in a heterogeneous grassland, a behavioural trade-off be- tween foraging and sheltering results in a demographic trade-off between growth and survivorship rates respectively, the balance of which is deter- mined by the patchy distribution of dicotyledonous plant leaf litter (Hassall, 1996). Animals exploiting patchy environments are known to modify their search path according to the quality of each patch (Gunn, 1937; Frankel & Gunn, 1961; Smith, 1974). Animals can modify their search path in a number of ways including modifying turning angle, frequency of turn alternation and walking speed (Stillman & Sutherland, 1990). This will consequently have effects on the amount of time an animal spends searching for high quality food and the amount of time it spends feeding on it.
We hypothesize that foraging behaviour of A. vulgare is affected by both food quality and the patchiness of high quality food resources. In this paper we therefore test the hypotheses that A. vulgare will alter its search path ac- cording to the quality of each patch in the grassland, and when high quality food is more clumped in space, so harder to Ž nd, it will spend less time feed- ing on it because it takes more time to Ž nd it. Experiments were designed to assess effects of spatial heterogeneity on foraging behaviour with treatments varying in the extent to which high quality food patches were clumped in space within a low quality food background. This represents in a simpliŽ ed way the variety in the patchiness of these food types in heterogeneous grass- lands (Hassall, 1996; Tuck, 2001).
Effect of food quality on growth
Growth chambers consisted of 170 £ 115 £ 55 mm transparent plastic boxes with lids con- taining a sloping base of plaster of Paris, sloping from a height of 30 mm at one end of the base to 10 mm at the other end, with 100 ml of water added to a foam rubber reservoir in the deeper end to maintain a humidity gradient along the length of the box. 50 ml of sand was placed over the base in the drier half of the box for shelter and an excess of 5.00 g of powdered food was placed on the plaster of Paris in the other half. One treatment contained S. olusatrum and the other F. ovina. There were Ž ve replicates of each treatment, each repli- cate box containing Ž ve individuals of differing sizes, over a comparable range in each box, so that specimens could be identiŽ ed individually. The experiment was run at 15±C (12 h light, 12 h dark). Food was replaced with the same amount after two weeks to reduce mould and dead individuals replaced by similar sized specimens. Individuals were weighed at the beginning, if they were replacements and after 30 days.
236 TUCK & HASSALL
Fig. 1. Design of experiment to investigate effects of spatial heterogeneity of high quality food on the foraging behaviour of Armadillidium vulgare.
Effects of spatial heterogeneity in food on foraging behaviour
Experimental arenas 600 £ 600 £ 80 mm were made from plastic gravel trays lined with 20 mm of plaster of Paris and covered with 620 £ 620 mm perspex airtight lids resting on foam rubber draft excluder Ž xed to the top of the tray walls to prevent animals escaping. The patchinessof a constant excess amount of high quality food, 1.90 g of powdered S. olusatrum, was altered in a low quality background of 9.60 g of powdered dead F. ovina. Foods were placed directly on the plaster of Paris in three treatments (Fig. 1): uniform, high quality food placed in alternate 75£75 mm squares with low quality food distributed in alternate squares; patchy,with the same quantity of high quality food restricted to nine 30 mm diameter patches; and clumped, with high quality food in one large 90 mm diameter patch. These arrangements of high quality food are within the range of patchiness of dicotyledonous plants observed in the Ž eld (Tuck, 2001).
The amount and patchiness of shelter was kept constant. The four shelters consisted of a 50 mm diameter petri dish half Ž lled with plaster of Paris (containing 8 ml of water) Ž lled to the brim with 10 ml of sand (with 3 ml of water added) to provide a refuge into which individuals could burrow during moulting, with a 52 mm diameter wooden circle with dowelling to support the centre (soaked in water) stuck into the sand. The edges of the shelters were 100 mm from the edge of the patchy and clumped food patches. To create a high relative humidity 200 ml of water was poured evenly on the plaster of Paris base in each arena before the start of each experiment. The observations were carried out in a dark room illuminated with a red light at approximately 20±C.
Ten medium sized (with headwidths of 0.8-1.1 mm) and ten large isopods (with head- widths of 1.4-2.0 mm) were given individual identities with ‘Tipp-Ex’ marks in different patterns on the pereion and placed in an arena. This density of 56 isopods m¡2 is comparable to densities found in the Ž eld (Hassall & DangerŽ eld, 1997). Isopods were acclimatized with
FORAGING BEHAVIOUR IN HETEROGENEOUS ENVIRONMENTS 237
12 h light at night time and 12 h dark during the day for one week before the experiment so that nocturnal foraging behaviour could be observed during the day. Experimental animals were acclimatized to the arena for 24 h before observations were made.
The detailed movements of 10 individual isopods selected in a random sequence were traced using an overhead projector pen on an acetate sheet for 15 minutes each, one after the other. All their movements and behavioural traits: time spent searching, turning frequency, turn angle, walking speed and time spent feeding in high and low quality food patches were recorded. Time spent searching was counted as time spent walking at any speed or in any direction whilst probing the surface with their antennae. Turning was counted as any turn that was made which was 10± or greater and feeding was identiŽ ed by movement of the mouthparts and head. This procedure was carried out for each treatment.
Data were analysed using two-way ANOVAs to Ž nd signiŽ cant effects of food quality on the different aspects of foraging behaviour and for signiŽ cant interactions between food quality and the patchinessof the treatment. For each aspect of foraging behaviour t -tests were carried out within each heterogeneity treatment to compare high and low quality food.
To analyse for signiŽ cant effects of the spatial heterogeneity of high quality food on the different aspects of foraging behaviour in both high and low quality food types one- way ANOVAs were used. Post-hoc tests were also carried out for each aspect of foraging behaviour in both types of food.
Effects of food quality on growth
The relative growth rate of A. vulgare (5:7 § 0:5 ¹g mg¡1 day¡1) was sig- niŽ cantly higher (t D 3:585, df D 8, p D 0:007) when fed a diet of exclu- sively S. olusatrum leaf litter than when fed a diet of only F. ovina leaf litter (1:9 § 0:9 ¹g mg¡1 day¡1).
Effects of food quality on foraging behaviour
In each treatment the amount of time spent in low and high quality food patches differed from the expected amount of time spent in each type of food patch by chance considering the area covered by each type of food, although this was only signiŽ cant for the patchy treatment (Table 1). In each treatment less time was spent in the low quality food backgrounds and more time in the high quality food patches than expected by chance.
Of the total time spent in a food type the mean percentage of time spent searching was signiŽ cantly higher in the low quality food background than in high quality food patches in both uniform and patchy treatments (Fig. 2a).
238 TUCK & HASSALL
¤ , w
FORAGING BEHAVIOUR IN HETEROGENEOUS ENVIRONMENTS 239
Fig. 2. Effects of ¤ low quality food Festuca ovina and high quality food Smyrnium olusatrum on the foraging behaviour of Armadillidium vulgare. (a) Mean proportion of time in each of low and high quality food patches that was spent searching as the high quality food becomes more clumped (effect of food quality; F1;41 D 6:691, p D 0:013). (b) Mean walking speed in low and high quality food patches as the high quality food becomes more clumped. (c) Mean number of turns cm¡1 in low and high quality food patches as the high quality food becomes more clumped (effect of food quality; F1;41 D 5:984, p D 0:019). (d) Mean turn angle in low and high quality food patches as the high quality food becomes more clumped (effect of food quality; F1;41 D 19:777, p < 0:001). (e) Mean proportion of time in each of low and high quality food patches that was spent feeding as the high quality food becomes more clumped (effect of food quality; F1;41 D 19:771, p < 0:001) (asterisks
between pairs of bars mean that the values differ at p < 0:05), N D 10.
240 TUCK & HASSALL
Isopods walked more slowly (Fig. 2b), turned more often (Fig. 2c) but at greater angles (Fig. 2d) in high quality food patches than in low quality ones, signiŽ cantly so in the patchy treatment.
A. vulgare spent a greater proportion of the total time that they were in a food type feeding when in high quality food patches than when in low quality food backgrounds in both uniform and patchy treatments (Fig. 2e).
Effects of spatial heterogeneity of high quality food on foraging behaviour
The proportion of time A. vulgare spent in an arena in the low quality food background increased signiŽ cantly and time in the high quality patches de- creased signiŽ cantly as the high quality food became more clumped (Fig. 3a & b). Increased patchiness of high quality food resulted in 61:0§ 13:7% less time spent in the clumped high quality patch than in uniform patches.
The turn frequency in the low quality background was lower in the patchy and clumped treatments than in the uniform treatment, signiŽ cantly so in the patchy treatment (Fig. 3c). The mean turn angle in high quality patches was signiŽ cantly higher in the patchy treatment than in the other two treatments (Fig. 3d).
There was no signiŽ cant effect of the spatial heterogeneity of high quality food on the mean proportion of time in low or high quality food patches spent searching or feeding or on the mean walking speed in low or high quality food patches.
It is evident that Armadillidium vulgare displays higher growth rates when fed a diet consisting solely of the umbellifer Smyrnium olusatrum litter com- pared to the grass Festuca ovina litter, which is likely to be a result of the sig- niŽ cantly higher nitrogen content of S. olusatrum litter (Tuck, 2001). Rush- ton & Hassall (1983) also showed that survivorship and reproduction were signiŽ cantly lower on a diet of F. ovina litter than when fed on litter from di- cotyledonous plants with higher nitrogen contents, the wider signiŽ cance of which was reviewed by Hassall & Rushton (1984). This experiment shows that F. ovina litter is a lower quality diet in that this isopod shows clear Ž t- ness costs when fed on it compared with when fed on S. olusatrum, which is therefore, by Crawley’s (1983) deŽ nition, a higher quality food.