Behavioural Processes, 26 (I992) 177-l 88 0 1992 ElsevierScience Ptiblishers B.V. All rights reserved 0376-6357/92/$05.00
BEPROC
177
00408
D.W. van Liere Department
of Animal Husbandry,
Ethology Section, Agricultural
University,
Wageningen,
The Netherlands
(Accepted 19 December 1991)
Abstract
After a prolonged experience with wood-shavings or with sand, 2 x 11 hens (Callus gal/us domesticus) were litter-deprived. Prior to the 7 to 8 day deprivation period feather samples were taken. This was repeated immediately after the deprivation had ended and right after the first bath in the familiar litter. Each feathe,, was cut at the transition between the proximal plumulous and distal pennaceous part and lipids were extracted from separate samples, containing one type of feather parts. Hens on wood-shavings bathed 10 minutes longer after deprivation than hens on sand. ‘This was due only to an extension of the second phase of the dustbath, which included rubbing; the first phase of tossing did not differ. In the extended phase of the wood-shavings bath the tendency to rub was lower, whereas the tendency to toss was higher than in the comparable phase of the sand bath. Thus, litter quality affected the amount of rubbings and of tossings in between. Rubbing did not effectuate a close contact between wood-shavings and the proximal integument, as wood-shavings could not be tossed into the plumage. This contrasted with baths in sand and only these baths resulted in a removal of excessive lipids from the plumulous parts. Therefore rubbing seems functionally crucial. The lipid level immediately after deprivation and the change in the quantity of lipids due to deprivation positively correlated with the amount of tossings in wood-shavings and with the amount of rubbings in sand. This indicates that the lipid condition is causally involved in dustbathing. Key words:
Dustbathing behaviour; Feather; Hen; Liter quality; Lipid; Lipid regulation; Integument
Correspondence
to: D.W. van Liere, Croenkampen
67, 3407 RK Assen, The Netherlands.
178
c”,tbathing in fowl is functionally organized in sequences of tossing and rubbing behaviour. Tossing, such as vertical wing shaking, is performed with the feathers fluffed and serves to distribute the litter over and into the plumage. Rubbing, which consists of side lying al>d side rubbing, is performed with the feathers flattened and the wings kept firmly to the body (cf. Kruijt, 1964; Borchelt, 1975; F&-h, 1981; Klinger, 1985; van Liere and Wiepkema, 1992). Rubbing is always preceded by tossing (in quail: Borchelt, 1975; in hens: van Liere et al., 1990; Vestergaard et al., 1990) and the litter that has penetrated into the plumage becomes enclosed by the flattening feathers. The contact between this litter and the proximal part of the integument may intensify as the hen presses her body against the rim of the dustbathing hole during side rubbing by stretching her legs (cf. van Liere and Wiepkema, 1992). Physical properties of the bathing material affect its penetration into the plumage: sand easily reaches up to the skin after tossing behaviour, whereas wood-shavings do not and adhere to the distal parts of the plumage (van _iere et al., 1990). The litter quality rna’f influence the composition of a bath, as tossingsseem to be more frequently reinitiated after rubbing in wood-shavings than in sand (van Liere et al., 1990). Moreover, the litter quality may affect the condition of the plumage. Uropygial glalrd lipids and lipids which are produced during epidermal keratinization, adhere to the feathers (Lucas and Stettenheim 1972b; lshida et al., 1973; Hodges, 1974; Borchelt et al., 1979; Jacob and Ziswiler, 1982). These lipids become excessive when fowl are deprived of dust. Dust used during bathing adsorbs and removes excessive lipids from the feathers (Healy and Thomas, 1973; Borchelt and Duncan, 1974; van Liere and Bokma, 1987). As a consequence, the exact location of the dust particles can be essential to the lipid maintenance of the plumage. If this is the case, lipid reductioT.4at the proximal integumental level is expected to be insignificant when the bathing litter only reaches the distal level. The present study compares the composition of baths in sand with those in wood-shavings in relation to the efficiency of lipid removal from proximal and distal feather parts. A differentiation between the effects on breast and back feathers is thought appropriate, as oiling behaviour is mainly directed to the breast and hardly to the back (van Liere et al., 1991). Moreover, since breast feathers contact the litter during lying behavicur they may be affected in a different way. The correlation between the amount of bathing and the quantity of feather lipids is of causal interest. It has been hypothesized that dustbathing is progressively stimulated whenever the feather lipid quantity increases over a certain critical value (Borchelt et al., 1973; Levine et al., 1974). There have been three papers in which this so-called lipid rc@ation hypothesis was tested by experimentally manipulating the level of feather lipids. Borchelt et al. (1979) examined in quail whether an artificial increase in feather lipid quantity would enhance the tendency to dustbathe and whether birds without uropygial glands were inclined to dustbathe less than intact birds. In neither case did they find an effect on the level of dustbathing. Nargaard-Nielsen and Vestergaard (1981) put the hypothesis to the test using four uropygial gland extirpated hens, but found these birds to have an even higher tendency to dustbathe than intact birds. However, their statistical analysis did not consider the small sample size; in fact the data did not show a difference between both treatments. Dustbathing seemed again not to be controlled by feather lipid quantity which contradicted the lipid regulation hypothesis. Nevertheless, neither Borchelt et al. nor NPrrgaard-Nielsen and Vestergaard had considered the quality of the feather
179
lipids, which either had been experiment& ap~!h rd or had accumulated from the birds’ skin. Van Liere et al. (1991) tested the lipid reg&u~n hypothesis by applying either fresh or stale uropygial gland lipids to the feadters. Now a significant dustbathing effect was found, but only in the case of the application of stale lipids. In particular, the amount of rubbing behaviour had increased. It was concluded that an increase of stale feather lipids stimulated dustbathing. Feather lipids accumulate (Borchelt and Duncan, 1974; van Liere and Bokma, 1987; van Liere et al., 19901, while becoming stale during dust deprivation (cf. Simmons, 1964; van Liere et al., 1991). An increase proportion of stale lipids may therefore be one of the factors stimulating dustbathing behaviour after dust deprivation (cf. Borchelt et al., 1973; Borchelt, 1975; Vestergaard, 1982; van Liere and Bokma, 1987; van Liere et a!., 1990). If it is assumed that the formation of stale and causally relevant lipid components parallels the accumulation of lipids on the feathers, then the lipid level at the end of the deprivation is expected to correiate with the amo nt of subsequent dustbathing behaviour. Conzlations between feather lipid levels and dustbathing quantities following a short-term iitter deprivation are examined in this stuoy, while distinguishing between tossing and rubbing behaviour, respectively.
aterials and methods Animals and hous.‘ng Twenty-four ISA Brown Warren laying hens with intact beak; were obtained from a commercial dealer at the age of 18 weeks. From hatching onwards they had been reared in wire cages without any litter; moreover, the food had not been accessible for dustbathing. On their arrival the hens were housed on wood-shavings in a 260 x 240 X 220 cm3 (I X w X h) pen. At the age of 33 weeks the hens were wing tagged for individual identification and randomly divided into four groups of six birds. These groups were housed in 150 x 260 x 260 cm3 pens with wooden slatted floors and nest boxes. The nest boxes did not contain any litter. The pens were placed in a single row and the groups were visually isolated from each other. In the corner of each pen, a 60 X 60 x 12 cm3 tray was enclosed in a 63 x 63 x 50 cm3 wire cage which could be entered !qrough a lock;ible sliding door. The tray was filled with litter up to 10 cm in depth. The Irtier was renewed twice a week. Sand was used for litter for the first and third pen in tt-,e row and wood-shavings for the other two pens. A 14/10 light/dark cycle (lights on at 0600 h) was maintained. Food from a small feeding trough (which did not permit hens to Jse food for bathing) and water were provided ad lib. Air temperatures averaged 18°C (rang6 : 14-21°C) and the average relative air humidity was 55% (range: 36-75%). The experiment started after 4 weeks of acclimatization. During the acclimatization period the animals were observed two to three times a week for about one hour in the afternoon. It appeared that the hens entered the trays and only dustbathed in the provided litter. Experimental design The experiment covered ten days. Feathers from each hen were individually sampled in the afternoon of day one. When these feathers were collected it was not known whether the birds hali dustbathed shortly before or were about to dustbathe. All hens were deprived of dustbathing material from day two to day eight by locking the wire cage that contained the dust tray. On day nine three hens and on day ten the other hens of a group were permitted to dustbathe one after the other (see Behavioural observations below)
180
immediately following a second feather sampling. In the period between the observations of day nine and day ten the wire cage remained locked. In this setup the first post-deprivation dustbaths of all hens were observed between 1300 and 1700 h which suited the daily distribution of dustbathing behaviour (Vestergaard, 1982; Vestergaard et al., 1990). Feathers were sampled for a third time immediately after each hen had dustbathed. Behavioural observations After a second sample of the plumage had been taken a hen was put into the tray inside the wire cage which remained closed to the other hens of the group. In this way, each hen was temporarily kept isolated in the litter tray, while only visual and auditory contact was possible with the hens outside the enclosure. As soon as the hen had been left in the wire cage, the observations started. Behaviour was continuously recorded with an QS-3 handheld computer (Observational Systems Inc., Seattle, Washington). The start of a dustbath was defined as the first occurrence of vertical wing shaking; the end was marked by body/wing shaking or by non-bathing behaviour, if this lasted for more than 5 min. Interruptions of less than s min were considered to belong to the dustbath. Dustbathing behaviour was divided into two phases. Phase 1 was defined to last until the first occurrence of rubbing (see belowi and, thus, comprised only tossing (see below) and certain non-dustbathing behaviour. Phase 2 was the remaining part of the dustbath, which, by definition, included sequences of rubbing, as well as sequences of tossing ant! other behaviour. Tossing behaviour wris defined as a sequence of vertical wing shaking, bill raking, head rubbing, scratching with one leg or lying, all sharing the characteristics of a fluffed olumage. Altholrgh bill raking could also be performed while the plumage was not fluffed, it was classified as “tossing behaviour”, because it occurs within the loop of bill raking, scratching with one leg, head rubbing, vertical wing shaking and bill raking (cf. for quail: Borrhelt, 1975; for hens: Vestergaard, 1981; van Liere, 1991). Rubbing behaviour was defined as a sequent e of side lying or side rubbing which were both invariably performed with the feathers flattened and the wings held tightly to the body. A detailed description of the dustbathing elements is given in van Liere and Wiepkema (1992). The minimum duration of a certain behaviour to be reliably measured was one second. Sequences either of bill raking, vertical wing shaking or scratching, in which the movements succeeded each other within one second, were recorded as one bill rake, vertical wing shake or scratch, respectively.
Feather sampling and lipid extraction Before litter deprivation, after litter deprivation and immediately after the first dustbath following deprivation, two feather samples were taken: one from the breast region and one from the back region of the plumage. Feathers were cut off at their base at eight fixed locations per region, i.e. 5 feathers per location. The locations at the back region were dorsally at the posterior cervical, the interscapular and dorso pelvic tract; those at the breast region were ventrally at the posterior cervical and at the pectoral tract (Lucas and Stettenheim, 1972a). Each feather was mature, appalently clean, and included a proximal plumulous part and a distal pennaceous part. After sampling, the feathers were cut at the transition between the plumulous and pennaceous part, which was recognized by the colour and structure of the barbs (white plumules and brown pennae). Thus, per herl 3 (time of sampling) x 2 (region of the plumage) x 2 (feather part) samples (weighing 0.5-I .O $, containing 40 feather parts each, were obtained.
181
A Soxhlett cold extraction method (Anonymous, 1983) was applied in order to extract lipids from the feather part samples. After two hours the solvent’(petroleum spirit: boiling range 40-60°C) containing the lipids was poured from the distillation receiver into a pot of glass, which Gighed 12 g. Fresh solvent was used twice to clean the receiver and, subsequently, added to the pot. After the solvent was distilled the pot, containing the lipid residue, was cooled down to room temperature and weighed. The weight of the lipid residue ranged between 2.8 and 11.6 mg. For the purpose of reliably estimating the weight of this residue, the 12 g pot was better suited than the distillation receiver, which weighed 60 g. The quantity of extracted lipids was expressed per gram of (non-dried) feathers. Statistics The total duration of the dustbaths was divided into the duration of phase 1 and the duration of phase 2 (cf. the definition in Behavioural observations). Per hen, the median duration of each dustbathing element and its total nu&n . .. ..r was iaiculated. Subsequently, the number of vertical wing shakes in phase 1 (representative for tossing) and of side rubbings in phase 2 (representative for rubbing) were expressed per minute in phase 1 and phase 2, respectively. The number of vertical wing shakes in phase 2 were expressed per minute in the phase 2 residue. The phase 2 residue was defined as the total duration of the intervals in between the rubbings. Median duration and total number of sequences, which included either only tossing, rubbing or rather behadiour (cf. Behavmml observations) were calculated per hen. These sequences were defined as tossing, rubbing and other behaviour events, respectively. Medians and third quartile deviations (the median - third quartile range) were calculated per substrate treatment for each of the aforementioned parameters. The median quantity of feather lipids was calculated per part of the feather, region of the plumage, sampling period and substrate treatment. A Wilcoxon independent sample test was used to compare between both substrate treatments. A Wilcoxon signed-ranked matched-pair test was used to compare within substrate treatments, regions of the plumage and parts of the feather (Conover, 1980). Relations between the quantity of feather lipids and the total time spent tossing or the total time spent rubbing and relations between the lipid levels of the different feather parts were examined using a Spearman ranking correlation (Cono\ver 1980) per substrate treatment. One hen in the sand treatment dustbathed for 12 min, but did not perform any rubbing behaviour. One hen in the wood-shavings treatment had injured her leg. These hens were excluded from the experiment which, as a consequence, reduced the sample c;ze to 11 hens per substrate treatment. All statistical tests were two-tailed.
Results The latency to start dustbathing after the 7 to 8 days of deprivation did not differ significantly between the substrate treatments (Table 1). The total duration of the dustbaths, however, was about 10 min longer on wood-shavings than on sand (Tab!e 1; P < 0.01). This difference was mainly caused by the duration of phase 2 (Table 1; P < 0.05). The durations of the dustbathing elements did not differ between both substrate treatments. Head rubbing and side rubbing lasted about 1 s, vertical wing shaking and scratching with one leg lasted about 2 s and bill raking and lying lasted about 3 s. Side lying
182 TABLE 1 Several parameters of bathing behaviour on sand or on wood-shavings after a 7 to 8 day litter deprivation. Medians (With third quartile deviations) are given; alI duration estimates in minutes In = 11 per treatment). Wood-shavings
Sand
Total duration
29.9iS.6)
2.1 (1.6) 39.0 (5.7)**
Duration phase 1
14.4 (4.1) 15.4 (3.4)
12.2t11.71 24.5 (6.3)*
2.3 (0.7)
Latency
Duration phase 2 Nr. of v.w.s./min.
phase 1
Nr. of v.w.s.jsllin.
phase 2 res.
Nr. of side r.,,‘min. phase 2 Duration tossing event Duration rubbing event Duration non-dustb. interruption Nr. of tossing events Nr. of rubbing events Nr. of non-dustb. interruption Sand-wood-shavings
2.2 (0.6)“ 1.3(0,2)' 2.7 (1.6) 1.8t4.1) 0.9 (1.0) 0.1 (0.4) 11 (4) 10 (4) 1 (1)
comparisons: (*): 0.05 c P < 0.10;
wood-shavings comparisons:
‘-‘: P < 0.01.
2.0 (0.51d 1.2 (0.3)‘ 1.2 (1.3)** 1.5 (2.4) 0.5 (0.2)** 0.6 (2.2)'*' 15 (lo)* 13 (8)'*' 2
(1)
*: P < 0.05;**: P < 0.01;
Within
sand or
Phase 1: phase from start until first rubbing behaviour;
Phase 2: phase from start of rubbing behaviour until end of dustbath; Phase 2 res.: phase 2 residue, i.e. phase 2 excluding the time spent on rubbing behaviour; v w.s.: vertical wing shake; side r.: side rubbing; non-dustb.:
non-dustbathing.
lasted significantly less in wood-shavings than in sand (8 (I 1 v. 11 (3) seconds; P < 0.05). The total numbers of bill raking, scratching with one leg, vertical wing shaking and lying, however, were significantly higher in L&s on wood-shavirigs compared to those on sand (Fig. 1). The total numbers of other bathing elements did not differ between the substrate treatments. When expressing the number of vertical wing shakes per min of phase 1 or per min in the phase 2 residue no significant differences were found between both substrate treatments. However, side rubbing was less frequently expressed per min in phase 2 in wood-shavings than in sand (Table 1; P < 0.01). Within both substrate treatments, the relative number of vertical wing shakes in the phase 2 residue was significantly lower than in the phase 1 (Table 1; P < 0.01 in both treatments). The duration of tossing events did not differ between the bathing substrates, whereas the duration of the rubbing events was significantly less in wood-shavings than in sand (Table 1; P < 0.01). The number of tossing events was significantly higher in wood-shavings than in sand (P < 0.051, while the number .Jf rubbing events tended to be higher too (Table 1; P = 0.09). The duration of the non-dustbathing events tended to be longer in wood-shavings compared to those in sand (Table 1; P = O.O7), whereas their number did not significantly differ. The quantity of lipids of the plumulous breast feather parts and of the plumulous back feather parts were significantly higher in the wood-shavings treatment than in the sand treatment prior to the deprivation of litter and after the first dustbath following deprivation
183
200
.*7
V. W. SHAKE
smA-TcH
BlLLRhKE 200
175
i75
150
150
125
125
100
100
l-0-l 100
LIE
r*,
h-0-l 300 250
75
200 50
150
75
100
50
25
25
50
0
0
SIDE LIE
0
SIDE RUB
50
SAND n
25
0
WOOD-
SHAVINGS
TREATMENT Fig. 1. Median total number (with third quartile deviation) of the elements in the first dustbath after a 7 to 8 day period of deprivation of sand or wood-shavings (scratch: scratch with one leg; V.W. shake: vertical wing shake). *: P c 0.05; * *: P < 0.01; n = 11 per substrate treatment.
(Fig. 2). However, these quantities did not significantly differ between subsi;ates immediately after deprivation (Fig. 2). The lipid level of the pennaceous parts of the feathers did not show any significant differences between substrates, irrespective of sampling period or of region of the plumage (Fig. 2). No significant differences in lipid levels were found between the different sampling periods in the wood-shavings treatment, except for the pennaceous parts of the breast. Their lipid levels tended to be highest immediately after deprivation, whereas the levels before deprivation and after dustbathing did not differ. In the sand treatment, bpwever, the deprivation significantly enhanced, and subsequent bathing significantly reduced, the level of lipids of the plumulous, as well as of the pennaceous parts of the breast feathers (Fig. 2). No significant differences were found between the samples prior to the deprivation and those after the first bath in sand. The plumulous parts of the back feathers also showed a significantly higher level of lipids due to the deprivation, whereas their lipid levels tended to be lower after the first dustbath, following deprivation (Fig. 2). No significant deprivation or bathing effect was found for the liplti level of the pennaceous parts of the back feathers. During the sampling of feathers after the first dustbath following deprivation, it was clear that the sand had penetrated the plumage during the bath, as some sand was still present on the skin of all sand bathing hens. Some sand particles still adhered to the proximal feather parts, too. The proximal feather parts of all hens that bathed in wood-shavings did not show any litter particles. The same was true of the skin of the hens kept on wood-shavings, except for the featherless tracts of the skin, such as the uropygial eminence and the adjacent region (cf. Lucas and Stettenheim, 1972a). In these hens wood-shavings did still adhere to the distal feather parts.
184
BREAST FUJU_OUS
BACK
FEATHERS PART
PENVACEWS
FEATHERS
!=~~hruoUSPART
PART
16 .
[*$*)I
PIFNACEOUS 16
14
14
12
12
10 .
PART
IO
T
16 14 12
10
IO
BEFORE
DEWIVATION
6
0
6
6
4
4
2
2
0
0
SAMPLING
PERIOD
AFTER
DEPRIVATION
#
MER
DJsrBATl-l
Fig. 2. Median quantity of lipids (with third quartile deviation) on plumulous and pennaceous parts of breast and back feathers that were sampled before deprivation, immediately after deprivation and after the first dustbath following the deprivation of sand or wood-shavings. (* ): 0.05 < P < 0.10; *: P 0.10; n = 11 per substrate treatment.
In wood-shavings,
of the amount of tossing behaviour nor with the amount of rubbing behaviour. Such correlations were also insignificant with respect to the feather lipd quantity after the first dustbath following deprivation, as well as with respect to the char,ge in feather lipid quantity due to this dustbath. The only exception was a positive correlation between the lipid level of the plumulous parts of the back feathers after the first bath and the amount of tossing behaviour in wood-shavings (rs = 0.61; P < 0.05). The lipid level immediately after deprivation and the change in lipid quantity due to the deprivation di show significant, positive correlations: in wood-shavings only with the amount of tossing ehaviour and in sand only with the amount of rubbing behaviour. The total time spent ssing in wood-shavings correlated with the quantity of lipids on the plumulous parts of t e back feathers immediately after deprivation, as well as with the quantitative change in lipids of these parts due to the deprivation (r, = 0.63 and 0.61; P =C0.05 in both cases). The time spent tossing in wood-shavings also correlated with the change in lipids on the pennaceous breast feather parts due to the deprivation (r, = 0.64; P < 0.05). In sand the total time spent rubbing positively related or tended to positively relate to the lipid level of the plumulous parts of the breast feathers, to the lipid level of the pennaceous parts of the breast feathers, and to the lipid level of the pennaceous parts of the back feat ers all sampled immediately after deprivation (r, = 0.62, 0.56 and 0.54; P < 0.05, P = 0.07 and P = 0.08, respectively). The total time spent lipids
before
as well
deprivation
did
as in sand, and for all types of feather parts the quantity not significantly
correlate
with
185
rubbing also positively correlated with the quantitative change in lipids on the pennaceous parts of the breast feathers due to the deprivation (rS = 0.80; P < 0.01). In the wood-shavings treatment the level of lipids on the plumulous breast feather parts correlated with that of the plumulous back feather parts immediately after deprivation, as well as after the first dustbath, following deprivation (rS = 0.82 and 0.61; P < 0.01 and P < 0.05, respectively). The lipid level of the plumulous breast feather parts correlated with that of the pennaceous breast feather parts for both c=mn JX...r!.rEi3, r._CE._.ztoo (r, = 0.90 and ina norinds 0.74; P < 0.001 and P < 0.01, respectively). In the sand treatment the lipid level of the plumulous breast feather parts correlated with that of the plumulous bact feather parts before deprivation, as well as after the first dustbath following deprivation Cr, = 0.63 and 0.84; P < 0.05 and P < 0.001, respectively). Further, no significant correlations per sampling period were found between the lipid levels of the different feather parts within the regions of the plumage or between the regions of the plumage withln the feather parts.
iscussion After a deprivation of 7 to 8 days, hens took longer baths in wood-shavings than in sand. This was due to an extension of the bath only after the start of rubbing behaviour (Table 1). However, the total number of side lying and side rubbing elements did not differ between both types of baths (Fig. I). The duration of the rubbing events, in particular the duration of the side lying element, and the number of the side rubbing elements per min in phase 2 did differ: the durations were shorter and the number was lower in wood-shavings than in sand (Table I!. The tendency to rub in wood-shavings was therefore reduced. This is presumably related to the lack of particles at the proximal integumental level. The stimulation of the proximal integument by a type of litter, which does reach between the feathers is presumed to be promoted when the hen holds her feathers and wings tightly to her body and rubs her body. In this view the results suggest that the feed-back during rubbing behaviour is crucial for the bathing programme. Tossing behaviour was not affected by the quality of the litter, because the amount of tossings and the duration of the bathing phase before the first rubbing did not differ between the substrate treatments (Table I). Moreover, the tendency to toss (vertical wing shake) per time unit in the phase 2 residue did not differ between the substrate treatments (Table I). The total number of tossing events and the total number of tossing elements, however, were higher in wood-shavings than in sand (Table 1; Fig. I). This implies that the tossing behdviour in between the rubbing behaviour occurred more often in wood-shavings than in sand. Although these features have to be verified for other types of litter and for hens that have experienced litter from hatching onwards, it is concluded that the litter quality primarily affects the facilitation of rubbing. The experience during rubbing behaviour seems to add to the causality of tossing: when the stimulation of rubbing is adequate, it inhibits tossing behaviour. If the dustbathing sequence is considered to be a chain of appetitive tossing and consummatory rubbing events, a new light may be shed on dust deprivation effects. For instance, the performance of several tossing elements and the side rubbing element is enhanced as a consequence of dust deprivation (in quail: Borchelt, 1975; verified in our lab with hens), but the effect on their mutual contingencies is not known. In addition, a shift developed in the patterning of tossing and rubbing in sham-dustbaths during long-term deprivation (van Liere and Wiepkema, 1992). This contrasted Vestergaard et al.‘s conclusion (1990) that the organization developed normally
186
for hens kept in a dustless environment. However, Vestergaard et al., did find that the hens took longer sham-dustbaths, while the frequency of side rubbing was lower compared to hens kept on sand. As a result of these findings Vestergaard et al.3 conclusion may have to be reconsidered: there might have been a significant effect on the alteration between rubbing and tossing as well. The difference in effect between wood-shavings and sand baths corresponded clearly to the exact location of the litter during its close integumental contact, as provided by rubbing behaviour. Wood-shavings did not reach the proximal plumulous parts of the integument, as a result of which lipids could not be removed from these parts. In contrast, sand easily reached between the feathers and excessive lipids from the plumulous parts were removed (Fig. 2). Moreover, a period of wood-shavings deprivation following ad lib. bathing did not affect the levei of lipids on the plumulous parts, whereas 2 period of sand deprivation did (Fig. 2). These results give reason to investigate whether functionalities concerning the proximal integument are affected by litter quality. Excessive lipids on the plumulous parts may cause the plumules to stick together which reduces the fluffiness of the down, and, consequently, the thermo-insulation of the plumage (van Liere and Bokma, 1987; van Liere and Siard, 1991). A pilot sampling of the back skin of hens that had been housed either on wood-shavings or on sand showed some, though non-significant, difference in thickness of the epidermal lipid layer. The epidermis had been 0~0, fixated and the average thickness of the lipid layer (SD) was 7.3 pm (2.6) and 6.0 pm (1.3), respectively (n = 8 per treatment). Inefficiency to remove lipids from the proximal integument leaves them to become stale (Simmons, 1964; van Liere et al., 1991). Such lipids may attract different species of ectoparasites (cf. Zeman, 1988; van Liere, 19911, while a change in lipid quality may alter chances of bacterial or mycotic infection of the integument (Jacob and Ziswiler, 1982). Moreover, the lipids on the skin appeared to include fractions of the horny layer. Excess of such fractions may be irritating and dustbathing may function to remove it from the proximal integument (cf. Kruijt, 1964; Dow, 1988). A functional surplus by the rubbing component of a dustbath can be tested by comparing the integument after baths that included tossing only with the integument after baths that included tossing, as well as rubbing. The feather lipid condition appeared to be causally involved in the amount of dustbathing, since onlv the lipid levels of feather parts sampled immediately after the deprivation and/or the quantitative lipid changes due to the deprivation positively correlated with the amount of subsequent tossing or rubbing behaviour. Cqe exception concerned the level of proximal feather lipids after bathing in wood-shavings, but the same correlation was found immediately after deprivation. It can be suggested that the exception was due to the poor effect of the wood-shavings on proximal feather lipids. It is intriguing that in case of sand only the amount of rubbing and in case of wood-shavings only the amount of tossing were related to lipid parameters. The relation found for sand baths is in line with van Liere et al.‘s findings (1991), at least when the formation of stale and causally relevant components corresponds to the accumulation of lipids during deprivation. Van Liere et al. found that specifically rubbing behaviour (in sand) was stimulated by stale lipids on the feathers. The rubbing component of the bath seems central in maintenance of the feather lipid condition: rubbing is affected by the quantity of stale lipids and is essential to an efficient removal of excessive lipids. Basically, this lipid regulation of rubbing behaviour corresponds to the early lipid regulation hypothesis of dustbathing (Borchelt et al., 1973; Levine et al., 1974). However, the functional differentiation between tossing and rubbing components on the one hand and the causal
deprived
187
significance of the feather lipid quality on the other hand necessitate a refinement of the regulation model. Furthermore, dustbathing behaviour is not only regulated by peripheral factors such as feather lipid condition, but also by intrinsical factors (Vestergaard et al., 1990; Hogan et al., 1991; van Liere, 1991; van Liere and Wiepkema, 1992). In wood-shavings, rubbing may not be sufficiently reinforced, resulting in an enhanced expression of tossing (Fig. 1, Table 1). The correlation between the lipid condition and the amount of tossing in wood-shavings may therefore be a spin off from the underlying causality between rubbing a,ld tossing. The relation between the lipid condition and the amount of tossing may have been reinforced in the course of the experience, as bathing in this litter has some, be it marginal, effect (e.g. reduction of lipids of pennaceous feather parts: Fig. 2). Several factors may explain why not all feather lipid measurements after the dust deprivation correlated with the dustbathing quantities. Lipid quantity might have ambiguously interacted with lipid quality and/or with other related peripheral factors (e.g. fluffiness, skin irritation) in the different parts of the plumage. Moreover, the lipid level of the pennaceous feather parts of the breast increased due to the deprivation, whereas the lipid level of the pennaceous feather parts of the back did not increase. This was found for both substrate treatments (Fig. 2) and corresponds to lshida et al.‘s (1973) and van Liere et al.‘s (1991) findings. lshida et al. suggested that hens oil distal feather parts more than proximal feather parts. Van Liere et al. showed that hens oil the breast most of all regions of the plumage. Hens may therefore not equally appreciate lipid conditions of the plumage at different locations, The correlations between the lipid level of the one feather part and the lipid level of the other feather part were ambiguous, because they differed between the regions of the plumage, the times of feather sampling and the types of irtter. Therefore, detailed observations of oiling allocations are additionally necessary in order to fully understand the maintenance of proximal and distal feather lipids.
The valuable comments on earlier drafts of this paper by Professor Piet Wiepkema, the assistance of Thijs Roest, the helpful histological support by John Dulos and Professor L.P.M. Timmermans are greatly appreciated. The same is true for Kees Zoeter’s instructions concerning the English language.
eferences Anonymous, 1983. Animal feeding stuffs: Determination of diethyl ether extract. International Standard 5986, International Standard Organization, Geneva, 5 pp. Borchelt, P.L., 1975. The organization of dustbathing components in Bobwhite virginianus). Behaviour, 53: 217-237. Borchelt, P.L., Eyer, J. and McHenry Jr, D.S., 1973. Dustbathing in Bobwhite virginianus) as a function of dust deprivation. Behav. Biol., 8: 109-l 14. Borchelt, P.L. and Duncan, L., 1974. Dustbathing and feather lipid in Bobwhite virginianus). Condor, 76: 471-472. Borchelt, P.L., Hoffman, T., Hurrell, R.M. and McCarthy, R., 1979. Regulation of Japanese Quail (Coturnix coturnixh J. Comp. Physiol. Psychol., 93: 134-l 39.
quail (Colinus quail (Cohnus quail (Colinus dustbathing in
188 Conover, w.J., 1980. Practical Nonparametric Statistics. John Wiley and Sons, New York, 493 pp. Dow, D.D., I 988. Dusting and sunning by Australian brush-turkeys. Emu, 88: 47-48. Falsch, D.W., 1981. Das Verhalten von Legehennen in unterschiedlichen Haltungssystemen unter Berucksichtigung der Auf zuchtmethoden. In: D.W. Fijlsch and K. Vestcrgaard (Editors), Das Verhalten von Hiihnern, Birkhzuser Verlag, Basel, pp. 9-106. Healy, W.M. and Thomas, J.W., 1973. Effects of dusting on plumage of lapanese quail. Wilson Bull., 85: 442-448. Hodges, R.D., 1974. The histology of the fowl. Academic Press, London, pp. 648. Hogan, J.A., Honrado, G.I. and Vestergaard, K., 1991. Development of a behavior system: Dustbathing in the Burmese red junglefowl (Callus gal/us spacliceus): II. Internal factors. 1. Comp. Psychol., 105: 269-273. Ishida, K., Suzuki, T,, Kusuhara, S. and Yamaguchi, M., 1973. Influence of preen gland removal on the lipid over the plumage of roosters. Poultry Sci., 52: 83-87. Jacob, J. and Ziswiler, V., 1982. The uropygial gland. In: D.S. Farner, J.R. King and K.C. Parkes (Editors), Avian biology. vol. VI. Academic Press Inc., New York, pp. 199-324. Klinger, G., 1985. Untersuchungen zum Staub- oder Sandbade Verhalten von Hijhnern. lnstitut fiir Tierhaltung und Tierztichtting der Universit;it Hohenheim, PhD-thesis, pp. 119. Kruijt, J.P., 1964. Ontogeny of social behaviour in Burmese red jungle fowl (Callus &/us spadiceus). Behaviour Suppl., 12: l-201. Levine, R.L., Hunter, I.E. and Borchelt, P.L., 1974. Dustbathing as a regulatory mechanism. Bull. Math. Biol., 36: 545-553. Liere, D.W. van, 1991. Function and organization of dustbathing in laying hens. Agricultural University, Wageningen, PhD-thesis, pp. 123. Liere, D.W. van and Bokma, S., ‘i987. Short-term feather maintenance as a function of dust-bathing in laying hens. Appl. Anim. Behav. Sci., 18: 197-204. Liere, D.W. van, Kooijman, J. and Wiepkema, P.R., 1990. Dustbathing behaviour of laying hens as related to quality of dustbathing material. Appl. Anim. Behav. Sci., 26: 127- 141. Liere, D.W. van, Aggrey, S.E., Brouns, F.M.R. and Wiepkema, P.R., 1991. Oiling behaviour and the effect of lipids on dustbathing behaviour in laying hens Callus gal/us domesficus. Behav. Proc., 24: 71-81. Liere, D.W. van and Siard, N., 1991. Towards an understanding of litter bathing quality in hens. In: MC. Appleby, R.I. Horrell, J.C. Petherick and SM. Rutter (Editors), Applied animal behaviour: past, present and future. UFAW, Potters Bar, Herts, pp. 132-133. Liere, D.W. van and Wiepkema P.R., 1992. Effects of long term deprivation of sand on dustbathing behaviour in laying hens. Anim. Behav., 43. Lucas, A.M. and Stettenheim, D.R., 1972a. Avian Anatomy, Part I. Agricultural Handbook 362, U.S. Government Printing Office, Washington, pp. 750. Lucas, A.M. and Stettenheim, D.R., 1972b. Avian Anatomy, Part II. Agricultural Handbook 362, U.S. Government Printing Office, Washington, pp. 340. Norgaard-Nielsen, G. and Vestergaard, K., 1981. Dustbathing behaviour of uropygial gland extirpated domestic hens. Effects of dust deprivation. Acta Vet. Stand., 22: 118-I 28. Simmons, K.E.L., 1964. Feather maintenance. In: A.L. Thompson (Editor), A New Dictionary of Birds, McGraw-Hill, New York, pp. 278-286. Vestergaard K., 1981. Behavioural and physiological studies of hens on wire floors and in deep litter pens. In: D.W. F&h and K. Vestergaard (Editors), Das Verhalten von Hiihnern, Birkh;iuser Verlag, Basel, pp. 115-l 32. V-tergaad
6 1982. Dust-bathing in the domestic fowl; diurnal rhythm and dust-deprivation. Anim. Ethol., 8: 487-495.
Vestergaard,
K., Hogan,
J.A. and Kruijt, J.P., 1990.
Appl.
The development of a behavior system: red junglefowl. I. The influence of the rearing environment on the organization of dustbathing. Behaviour, 1 I 2: 99-I I 6. Zeman, P., 1988. Surface skin lipids of birds- a proper host kariomone and feeding inducer in the Poultry red mite (Demanysws galhe). Exp. Appl. Acarol., 5: 163-l 73. Dustbathing
in the hmnese
BEPROC
177
00408
D.W. van Liere Department
of Animal Husbandry,
Ethology Section, Agricultural
University,
Wageningen,
The Netherlands
(Accepted 19 December 1991)
Abstract
After a prolonged experience with wood-shavings or with sand, 2 x 11 hens (Callus gal/us domesticus) were litter-deprived. Prior to the 7 to 8 day deprivation period feather samples were taken. This was repeated immediately after the deprivation had ended and right after the first bath in the familiar litter. Each feathe,, was cut at the transition between the proximal plumulous and distal pennaceous part and lipids were extracted from separate samples, containing one type of feather parts. Hens on wood-shavings bathed 10 minutes longer after deprivation than hens on sand. ‘This was due only to an extension of the second phase of the dustbath, which included rubbing; the first phase of tossing did not differ. In the extended phase of the wood-shavings bath the tendency to rub was lower, whereas the tendency to toss was higher than in the comparable phase of the sand bath. Thus, litter quality affected the amount of rubbings and of tossings in between. Rubbing did not effectuate a close contact between wood-shavings and the proximal integument, as wood-shavings could not be tossed into the plumage. This contrasted with baths in sand and only these baths resulted in a removal of excessive lipids from the plumulous parts. Therefore rubbing seems functionally crucial. The lipid level immediately after deprivation and the change in the quantity of lipids due to deprivation positively correlated with the amount of tossings in wood-shavings and with the amount of rubbings in sand. This indicates that the lipid condition is causally involved in dustbathing. Key words:
Dustbathing behaviour; Feather; Hen; Liter quality; Lipid; Lipid regulation; Integument
Correspondence
to: D.W. van Liere, Croenkampen
67, 3407 RK Assen, The Netherlands.
178
c”,tbathing in fowl is functionally organized in sequences of tossing and rubbing behaviour. Tossing, such as vertical wing shaking, is performed with the feathers fluffed and serves to distribute the litter over and into the plumage. Rubbing, which consists of side lying al>d side rubbing, is performed with the feathers flattened and the wings kept firmly to the body (cf. Kruijt, 1964; Borchelt, 1975; F&-h, 1981; Klinger, 1985; van Liere and Wiepkema, 1992). Rubbing is always preceded by tossing (in quail: Borchelt, 1975; in hens: van Liere et al., 1990; Vestergaard et al., 1990) and the litter that has penetrated into the plumage becomes enclosed by the flattening feathers. The contact between this litter and the proximal part of the integument may intensify as the hen presses her body against the rim of the dustbathing hole during side rubbing by stretching her legs (cf. van Liere and Wiepkema, 1992). Physical properties of the bathing material affect its penetration into the plumage: sand easily reaches up to the skin after tossing behaviour, whereas wood-shavings do not and adhere to the distal parts of the plumage (van _iere et al., 1990). The litter quality rna’f influence the composition of a bath, as tossingsseem to be more frequently reinitiated after rubbing in wood-shavings than in sand (van Liere et al., 1990). Moreover, the litter quality may affect the condition of the plumage. Uropygial glalrd lipids and lipids which are produced during epidermal keratinization, adhere to the feathers (Lucas and Stettenheim 1972b; lshida et al., 1973; Hodges, 1974; Borchelt et al., 1979; Jacob and Ziswiler, 1982). These lipids become excessive when fowl are deprived of dust. Dust used during bathing adsorbs and removes excessive lipids from the feathers (Healy and Thomas, 1973; Borchelt and Duncan, 1974; van Liere and Bokma, 1987). As a consequence, the exact location of the dust particles can be essential to the lipid maintenance of the plumage. If this is the case, lipid reductioT.4at the proximal integumental level is expected to be insignificant when the bathing litter only reaches the distal level. The present study compares the composition of baths in sand with those in wood-shavings in relation to the efficiency of lipid removal from proximal and distal feather parts. A differentiation between the effects on breast and back feathers is thought appropriate, as oiling behaviour is mainly directed to the breast and hardly to the back (van Liere et al., 1991). Moreover, since breast feathers contact the litter during lying behavicur they may be affected in a different way. The correlation between the amount of bathing and the quantity of feather lipids is of causal interest. It has been hypothesized that dustbathing is progressively stimulated whenever the feather lipid quantity increases over a certain critical value (Borchelt et al., 1973; Levine et al., 1974). There have been three papers in which this so-called lipid rc@ation hypothesis was tested by experimentally manipulating the level of feather lipids. Borchelt et al. (1979) examined in quail whether an artificial increase in feather lipid quantity would enhance the tendency to dustbathe and whether birds without uropygial glands were inclined to dustbathe less than intact birds. In neither case did they find an effect on the level of dustbathing. Nargaard-Nielsen and Vestergaard (1981) put the hypothesis to the test using four uropygial gland extirpated hens, but found these birds to have an even higher tendency to dustbathe than intact birds. However, their statistical analysis did not consider the small sample size; in fact the data did not show a difference between both treatments. Dustbathing seemed again not to be controlled by feather lipid quantity which contradicted the lipid regulation hypothesis. Nevertheless, neither Borchelt et al. nor NPrrgaard-Nielsen and Vestergaard had considered the quality of the feather
179
lipids, which either had been experiment& ap~!h rd or had accumulated from the birds’ skin. Van Liere et al. (1991) tested the lipid reg&u~n hypothesis by applying either fresh or stale uropygial gland lipids to the feadters. Now a significant dustbathing effect was found, but only in the case of the application of stale lipids. In particular, the amount of rubbing behaviour had increased. It was concluded that an increase of stale feather lipids stimulated dustbathing. Feather lipids accumulate (Borchelt and Duncan, 1974; van Liere and Bokma, 1987; van Liere et al., 19901, while becoming stale during dust deprivation (cf. Simmons, 1964; van Liere et al., 1991). An increase proportion of stale lipids may therefore be one of the factors stimulating dustbathing behaviour after dust deprivation (cf. Borchelt et al., 1973; Borchelt, 1975; Vestergaard, 1982; van Liere and Bokma, 1987; van Liere et a!., 1990). If it is assumed that the formation of stale and causally relevant lipid components parallels the accumulation of lipids on the feathers, then the lipid level at the end of the deprivation is expected to correiate with the amo nt of subsequent dustbathing behaviour. Conzlations between feather lipid levels and dustbathing quantities following a short-term iitter deprivation are examined in this stuoy, while distinguishing between tossing and rubbing behaviour, respectively.
aterials and methods Animals and hous.‘ng Twenty-four ISA Brown Warren laying hens with intact beak; were obtained from a commercial dealer at the age of 18 weeks. From hatching onwards they had been reared in wire cages without any litter; moreover, the food had not been accessible for dustbathing. On their arrival the hens were housed on wood-shavings in a 260 x 240 X 220 cm3 (I X w X h) pen. At the age of 33 weeks the hens were wing tagged for individual identification and randomly divided into four groups of six birds. These groups were housed in 150 x 260 x 260 cm3 pens with wooden slatted floors and nest boxes. The nest boxes did not contain any litter. The pens were placed in a single row and the groups were visually isolated from each other. In the corner of each pen, a 60 X 60 x 12 cm3 tray was enclosed in a 63 x 63 x 50 cm3 wire cage which could be entered !qrough a lock;ible sliding door. The tray was filled with litter up to 10 cm in depth. The Irtier was renewed twice a week. Sand was used for litter for the first and third pen in tt-,e row and wood-shavings for the other two pens. A 14/10 light/dark cycle (lights on at 0600 h) was maintained. Food from a small feeding trough (which did not permit hens to Jse food for bathing) and water were provided ad lib. Air temperatures averaged 18°C (rang6 : 14-21°C) and the average relative air humidity was 55% (range: 36-75%). The experiment started after 4 weeks of acclimatization. During the acclimatization period the animals were observed two to three times a week for about one hour in the afternoon. It appeared that the hens entered the trays and only dustbathed in the provided litter. Experimental design The experiment covered ten days. Feathers from each hen were individually sampled in the afternoon of day one. When these feathers were collected it was not known whether the birds hali dustbathed shortly before or were about to dustbathe. All hens were deprived of dustbathing material from day two to day eight by locking the wire cage that contained the dust tray. On day nine three hens and on day ten the other hens of a group were permitted to dustbathe one after the other (see Behavioural observations below)
180
immediately following a second feather sampling. In the period between the observations of day nine and day ten the wire cage remained locked. In this setup the first post-deprivation dustbaths of all hens were observed between 1300 and 1700 h which suited the daily distribution of dustbathing behaviour (Vestergaard, 1982; Vestergaard et al., 1990). Feathers were sampled for a third time immediately after each hen had dustbathed. Behavioural observations After a second sample of the plumage had been taken a hen was put into the tray inside the wire cage which remained closed to the other hens of the group. In this way, each hen was temporarily kept isolated in the litter tray, while only visual and auditory contact was possible with the hens outside the enclosure. As soon as the hen had been left in the wire cage, the observations started. Behaviour was continuously recorded with an QS-3 handheld computer (Observational Systems Inc., Seattle, Washington). The start of a dustbath was defined as the first occurrence of vertical wing shaking; the end was marked by body/wing shaking or by non-bathing behaviour, if this lasted for more than 5 min. Interruptions of less than s min were considered to belong to the dustbath. Dustbathing behaviour was divided into two phases. Phase 1 was defined to last until the first occurrence of rubbing (see belowi and, thus, comprised only tossing (see below) and certain non-dustbathing behaviour. Phase 2 was the remaining part of the dustbath, which, by definition, included sequences of rubbing, as well as sequences of tossing ant! other behaviour. Tossing behaviour wris defined as a sequence of vertical wing shaking, bill raking, head rubbing, scratching with one leg or lying, all sharing the characteristics of a fluffed olumage. Altholrgh bill raking could also be performed while the plumage was not fluffed, it was classified as “tossing behaviour”, because it occurs within the loop of bill raking, scratching with one leg, head rubbing, vertical wing shaking and bill raking (cf. for quail: Borrhelt, 1975; for hens: Vestergaard, 1981; van Liere, 1991). Rubbing behaviour was defined as a sequent e of side lying or side rubbing which were both invariably performed with the feathers flattened and the wings held tightly to the body. A detailed description of the dustbathing elements is given in van Liere and Wiepkema (1992). The minimum duration of a certain behaviour to be reliably measured was one second. Sequences either of bill raking, vertical wing shaking or scratching, in which the movements succeeded each other within one second, were recorded as one bill rake, vertical wing shake or scratch, respectively.
Feather sampling and lipid extraction Before litter deprivation, after litter deprivation and immediately after the first dustbath following deprivation, two feather samples were taken: one from the breast region and one from the back region of the plumage. Feathers were cut off at their base at eight fixed locations per region, i.e. 5 feathers per location. The locations at the back region were dorsally at the posterior cervical, the interscapular and dorso pelvic tract; those at the breast region were ventrally at the posterior cervical and at the pectoral tract (Lucas and Stettenheim, 1972a). Each feather was mature, appalently clean, and included a proximal plumulous part and a distal pennaceous part. After sampling, the feathers were cut at the transition between the plumulous and pennaceous part, which was recognized by the colour and structure of the barbs (white plumules and brown pennae). Thus, per herl 3 (time of sampling) x 2 (region of the plumage) x 2 (feather part) samples (weighing 0.5-I .O $, containing 40 feather parts each, were obtained.
181
A Soxhlett cold extraction method (Anonymous, 1983) was applied in order to extract lipids from the feather part samples. After two hours the solvent’(petroleum spirit: boiling range 40-60°C) containing the lipids was poured from the distillation receiver into a pot of glass, which Gighed 12 g. Fresh solvent was used twice to clean the receiver and, subsequently, added to the pot. After the solvent was distilled the pot, containing the lipid residue, was cooled down to room temperature and weighed. The weight of the lipid residue ranged between 2.8 and 11.6 mg. For the purpose of reliably estimating the weight of this residue, the 12 g pot was better suited than the distillation receiver, which weighed 60 g. The quantity of extracted lipids was expressed per gram of (non-dried) feathers. Statistics The total duration of the dustbaths was divided into the duration of phase 1 and the duration of phase 2 (cf. the definition in Behavioural observations). Per hen, the median duration of each dustbathing element and its total nu&n . .. ..r was iaiculated. Subsequently, the number of vertical wing shakes in phase 1 (representative for tossing) and of side rubbings in phase 2 (representative for rubbing) were expressed per minute in phase 1 and phase 2, respectively. The number of vertical wing shakes in phase 2 were expressed per minute in the phase 2 residue. The phase 2 residue was defined as the total duration of the intervals in between the rubbings. Median duration and total number of sequences, which included either only tossing, rubbing or rather behadiour (cf. Behavmml observations) were calculated per hen. These sequences were defined as tossing, rubbing and other behaviour events, respectively. Medians and third quartile deviations (the median - third quartile range) were calculated per substrate treatment for each of the aforementioned parameters. The median quantity of feather lipids was calculated per part of the feather, region of the plumage, sampling period and substrate treatment. A Wilcoxon independent sample test was used to compare between both substrate treatments. A Wilcoxon signed-ranked matched-pair test was used to compare within substrate treatments, regions of the plumage and parts of the feather (Conover, 1980). Relations between the quantity of feather lipids and the total time spent tossing or the total time spent rubbing and relations between the lipid levels of the different feather parts were examined using a Spearman ranking correlation (Cono\ver 1980) per substrate treatment. One hen in the sand treatment dustbathed for 12 min, but did not perform any rubbing behaviour. One hen in the wood-shavings treatment had injured her leg. These hens were excluded from the experiment which, as a consequence, reduced the sample c;ze to 11 hens per substrate treatment. All statistical tests were two-tailed.
Results The latency to start dustbathing after the 7 to 8 days of deprivation did not differ significantly between the substrate treatments (Table 1). The total duration of the dustbaths, however, was about 10 min longer on wood-shavings than on sand (Tab!e 1; P < 0.01). This difference was mainly caused by the duration of phase 2 (Table 1; P < 0.05). The durations of the dustbathing elements did not differ between both substrate treatments. Head rubbing and side rubbing lasted about 1 s, vertical wing shaking and scratching with one leg lasted about 2 s and bill raking and lying lasted about 3 s. Side lying
182 TABLE 1 Several parameters of bathing behaviour on sand or on wood-shavings after a 7 to 8 day litter deprivation. Medians (With third quartile deviations) are given; alI duration estimates in minutes In = 11 per treatment). Wood-shavings
Sand
Total duration
29.9iS.6)
2.1 (1.6) 39.0 (5.7)**
Duration phase 1
14.4 (4.1) 15.4 (3.4)
12.2t11.71 24.5 (6.3)*
2.3 (0.7)
Latency
Duration phase 2 Nr. of v.w.s./min.
phase 1
Nr. of v.w.s.jsllin.
phase 2 res.
Nr. of side r.,,‘min. phase 2 Duration tossing event Duration rubbing event Duration non-dustb. interruption Nr. of tossing events Nr. of rubbing events Nr. of non-dustb. interruption Sand-wood-shavings
2.2 (0.6)“ 1.3(0,2)' 2.7 (1.6) 1.8t4.1) 0.9 (1.0) 0.1 (0.4) 11 (4) 10 (4) 1 (1)
comparisons: (*): 0.05 c P < 0.10;
wood-shavings comparisons:
‘-‘: P < 0.01.
2.0 (0.51d 1.2 (0.3)‘ 1.2 (1.3)** 1.5 (2.4) 0.5 (0.2)** 0.6 (2.2)'*' 15 (lo)* 13 (8)'*' 2
(1)
*: P < 0.05;**: P < 0.01;
Within
sand or
Phase 1: phase from start until first rubbing behaviour;
Phase 2: phase from start of rubbing behaviour until end of dustbath; Phase 2 res.: phase 2 residue, i.e. phase 2 excluding the time spent on rubbing behaviour; v w.s.: vertical wing shake; side r.: side rubbing; non-dustb.:
non-dustbathing.
lasted significantly less in wood-shavings than in sand (8 (I 1 v. 11 (3) seconds; P < 0.05). The total numbers of bill raking, scratching with one leg, vertical wing shaking and lying, however, were significantly higher in L&s on wood-shavirigs compared to those on sand (Fig. 1). The total numbers of other bathing elements did not differ between the substrate treatments. When expressing the number of vertical wing shakes per min of phase 1 or per min in the phase 2 residue no significant differences were found between both substrate treatments. However, side rubbing was less frequently expressed per min in phase 2 in wood-shavings than in sand (Table 1; P < 0.01). Within both substrate treatments, the relative number of vertical wing shakes in the phase 2 residue was significantly lower than in the phase 1 (Table 1; P < 0.01 in both treatments). The duration of tossing events did not differ between the bathing substrates, whereas the duration of the rubbing events was significantly less in wood-shavings than in sand (Table 1; P < 0.01). The number of tossing events was significantly higher in wood-shavings than in sand (P < 0.051, while the number .Jf rubbing events tended to be higher too (Table 1; P = 0.09). The duration of the non-dustbathing events tended to be longer in wood-shavings compared to those in sand (Table 1; P = O.O7), whereas their number did not significantly differ. The quantity of lipids of the plumulous breast feather parts and of the plumulous back feather parts were significantly higher in the wood-shavings treatment than in the sand treatment prior to the deprivation of litter and after the first dustbath following deprivation
183
200
.*7
V. W. SHAKE
smA-TcH
BlLLRhKE 200
175
i75
150
150
125
125
100
100
l-0-l 100
LIE
r*,
h-0-l 300 250
75
200 50
150
75
100
50
25
25
50
0
0
SIDE LIE
0
SIDE RUB
50
SAND n
25
0
WOOD-
SHAVINGS
TREATMENT Fig. 1. Median total number (with third quartile deviation) of the elements in the first dustbath after a 7 to 8 day period of deprivation of sand or wood-shavings (scratch: scratch with one leg; V.W. shake: vertical wing shake). *: P c 0.05; * *: P < 0.01; n = 11 per substrate treatment.
(Fig. 2). However, these quantities did not significantly differ between subsi;ates immediately after deprivation (Fig. 2). The lipid level of the pennaceous parts of the feathers did not show any significant differences between substrates, irrespective of sampling period or of region of the plumage (Fig. 2). No significant differences in lipid levels were found between the different sampling periods in the wood-shavings treatment, except for the pennaceous parts of the breast. Their lipid levels tended to be highest immediately after deprivation, whereas the levels before deprivation and after dustbathing did not differ. In the sand treatment, bpwever, the deprivation significantly enhanced, and subsequent bathing significantly reduced, the level of lipids of the plumulous, as well as of the pennaceous parts of the breast feathers (Fig. 2). No significant differences were found between the samples prior to the deprivation and those after the first bath in sand. The plumulous parts of the back feathers also showed a significantly higher level of lipids due to the deprivation, whereas their lipid levels tended to be lower after the first dustbath, following deprivation (Fig. 2). No significant deprivation or bathing effect was found for the liplti level of the pennaceous parts of the back feathers. During the sampling of feathers after the first dustbath following deprivation, it was clear that the sand had penetrated the plumage during the bath, as some sand was still present on the skin of all sand bathing hens. Some sand particles still adhered to the proximal feather parts, too. The proximal feather parts of all hens that bathed in wood-shavings did not show any litter particles. The same was true of the skin of the hens kept on wood-shavings, except for the featherless tracts of the skin, such as the uropygial eminence and the adjacent region (cf. Lucas and Stettenheim, 1972a). In these hens wood-shavings did still adhere to the distal feather parts.
184
BREAST FUJU_OUS
BACK
FEATHERS PART
PENVACEWS
FEATHERS
!=~~hruoUSPART
PART
16 .
[*$*)I
PIFNACEOUS 16
14
14
12
12
10 .
PART
IO
T
16 14 12
10
IO
BEFORE
DEWIVATION
6
0
6
6
4
4
2
2
0
0
SAMPLING
PERIOD
AFTER
DEPRIVATION
#
MER
DJsrBATl-l
Fig. 2. Median quantity of lipids (with third quartile deviation) on plumulous and pennaceous parts of breast and back feathers that were sampled before deprivation, immediately after deprivation and after the first dustbath following the deprivation of sand or wood-shavings. (* ): 0.05 < P < 0.10; *: P 0.10; n = 11 per substrate treatment.
In wood-shavings,
of the amount of tossing behaviour nor with the amount of rubbing behaviour. Such correlations were also insignificant with respect to the feather lipd quantity after the first dustbath following deprivation, as well as with respect to the char,ge in feather lipid quantity due to this dustbath. The only exception was a positive correlation between the lipid level of the plumulous parts of the back feathers after the first bath and the amount of tossing behaviour in wood-shavings (rs = 0.61; P < 0.05). The lipid level immediately after deprivation and the change in lipid quantity due to the deprivation di show significant, positive correlations: in wood-shavings only with the amount of tossing ehaviour and in sand only with the amount of rubbing behaviour. The total time spent ssing in wood-shavings correlated with the quantity of lipids on the plumulous parts of t e back feathers immediately after deprivation, as well as with the quantitative change in lipids of these parts due to the deprivation (r, = 0.63 and 0.61; P =C0.05 in both cases). The time spent tossing in wood-shavings also correlated with the change in lipids on the pennaceous breast feather parts due to the deprivation (r, = 0.64; P < 0.05). In sand the total time spent rubbing positively related or tended to positively relate to the lipid level of the plumulous parts of the breast feathers, to the lipid level of the pennaceous parts of the breast feathers, and to the lipid level of the pennaceous parts of the back feat ers all sampled immediately after deprivation (r, = 0.62, 0.56 and 0.54; P < 0.05, P = 0.07 and P = 0.08, respectively). The total time spent lipids
before
as well
deprivation
did
as in sand, and for all types of feather parts the quantity not significantly
correlate
with
185
rubbing also positively correlated with the quantitative change in lipids on the pennaceous parts of the breast feathers due to the deprivation (rS = 0.80; P < 0.01). In the wood-shavings treatment the level of lipids on the plumulous breast feather parts correlated with that of the plumulous back feather parts immediately after deprivation, as well as after the first dustbath, following deprivation (rS = 0.82 and 0.61; P < 0.01 and P < 0.05, respectively). The lipid level of the plumulous breast feather parts correlated with that of the pennaceous breast feather parts for both c=mn JX...r!.rEi3, r._CE._.ztoo (r, = 0.90 and ina norinds 0.74; P < 0.001 and P < 0.01, respectively). In the sand treatment the lipid level of the plumulous breast feather parts correlated with that of the plumulous bact feather parts before deprivation, as well as after the first dustbath following deprivation Cr, = 0.63 and 0.84; P < 0.05 and P < 0.001, respectively). Further, no significant correlations per sampling period were found between the lipid levels of the different feather parts within the regions of the plumage or between the regions of the plumage withln the feather parts.
iscussion After a deprivation of 7 to 8 days, hens took longer baths in wood-shavings than in sand. This was due to an extension of the bath only after the start of rubbing behaviour (Table 1). However, the total number of side lying and side rubbing elements did not differ between both types of baths (Fig. I). The duration of the rubbing events, in particular the duration of the side lying element, and the number of the side rubbing elements per min in phase 2 did differ: the durations were shorter and the number was lower in wood-shavings than in sand (Table I!. The tendency to rub in wood-shavings was therefore reduced. This is presumably related to the lack of particles at the proximal integumental level. The stimulation of the proximal integument by a type of litter, which does reach between the feathers is presumed to be promoted when the hen holds her feathers and wings tightly to her body and rubs her body. In this view the results suggest that the feed-back during rubbing behaviour is crucial for the bathing programme. Tossing behaviour was not affected by the quality of the litter, because the amount of tossings and the duration of the bathing phase before the first rubbing did not differ between the substrate treatments (Table I). Moreover, the tendency to toss (vertical wing shake) per time unit in the phase 2 residue did not differ between the substrate treatments (Table I). The total number of tossing events and the total number of tossing elements, however, were higher in wood-shavings than in sand (Table 1; Fig. I). This implies that the tossing behdviour in between the rubbing behaviour occurred more often in wood-shavings than in sand. Although these features have to be verified for other types of litter and for hens that have experienced litter from hatching onwards, it is concluded that the litter quality primarily affects the facilitation of rubbing. The experience during rubbing behaviour seems to add to the causality of tossing: when the stimulation of rubbing is adequate, it inhibits tossing behaviour. If the dustbathing sequence is considered to be a chain of appetitive tossing and consummatory rubbing events, a new light may be shed on dust deprivation effects. For instance, the performance of several tossing elements and the side rubbing element is enhanced as a consequence of dust deprivation (in quail: Borchelt, 1975; verified in our lab with hens), but the effect on their mutual contingencies is not known. In addition, a shift developed in the patterning of tossing and rubbing in sham-dustbaths during long-term deprivation (van Liere and Wiepkema, 1992). This contrasted Vestergaard et al.‘s conclusion (1990) that the organization developed normally
186
for hens kept in a dustless environment. However, Vestergaard et al., did find that the hens took longer sham-dustbaths, while the frequency of side rubbing was lower compared to hens kept on sand. As a result of these findings Vestergaard et al.3 conclusion may have to be reconsidered: there might have been a significant effect on the alteration between rubbing and tossing as well. The difference in effect between wood-shavings and sand baths corresponded clearly to the exact location of the litter during its close integumental contact, as provided by rubbing behaviour. Wood-shavings did not reach the proximal plumulous parts of the integument, as a result of which lipids could not be removed from these parts. In contrast, sand easily reached between the feathers and excessive lipids from the plumulous parts were removed (Fig. 2). Moreover, a period of wood-shavings deprivation following ad lib. bathing did not affect the levei of lipids on the plumulous parts, whereas 2 period of sand deprivation did (Fig. 2). These results give reason to investigate whether functionalities concerning the proximal integument are affected by litter quality. Excessive lipids on the plumulous parts may cause the plumules to stick together which reduces the fluffiness of the down, and, consequently, the thermo-insulation of the plumage (van Liere and Bokma, 1987; van Liere and Siard, 1991). A pilot sampling of the back skin of hens that had been housed either on wood-shavings or on sand showed some, though non-significant, difference in thickness of the epidermal lipid layer. The epidermis had been 0~0, fixated and the average thickness of the lipid layer (SD) was 7.3 pm (2.6) and 6.0 pm (1.3), respectively (n = 8 per treatment). Inefficiency to remove lipids from the proximal integument leaves them to become stale (Simmons, 1964; van Liere et al., 1991). Such lipids may attract different species of ectoparasites (cf. Zeman, 1988; van Liere, 19911, while a change in lipid quality may alter chances of bacterial or mycotic infection of the integument (Jacob and Ziswiler, 1982). Moreover, the lipids on the skin appeared to include fractions of the horny layer. Excess of such fractions may be irritating and dustbathing may function to remove it from the proximal integument (cf. Kruijt, 1964; Dow, 1988). A functional surplus by the rubbing component of a dustbath can be tested by comparing the integument after baths that included tossing only with the integument after baths that included tossing, as well as rubbing. The feather lipid condition appeared to be causally involved in the amount of dustbathing, since onlv the lipid levels of feather parts sampled immediately after the deprivation and/or the quantitative lipid changes due to the deprivation positively correlated with the amount of subsequent tossing or rubbing behaviour. Cqe exception concerned the level of proximal feather lipids after bathing in wood-shavings, but the same correlation was found immediately after deprivation. It can be suggested that the exception was due to the poor effect of the wood-shavings on proximal feather lipids. It is intriguing that in case of sand only the amount of rubbing and in case of wood-shavings only the amount of tossing were related to lipid parameters. The relation found for sand baths is in line with van Liere et al.‘s findings (1991), at least when the formation of stale and causally relevant components corresponds to the accumulation of lipids during deprivation. Van Liere et al. found that specifically rubbing behaviour (in sand) was stimulated by stale lipids on the feathers. The rubbing component of the bath seems central in maintenance of the feather lipid condition: rubbing is affected by the quantity of stale lipids and is essential to an efficient removal of excessive lipids. Basically, this lipid regulation of rubbing behaviour corresponds to the early lipid regulation hypothesis of dustbathing (Borchelt et al., 1973; Levine et al., 1974). However, the functional differentiation between tossing and rubbing components on the one hand and the causal
deprived
187
significance of the feather lipid quality on the other hand necessitate a refinement of the regulation model. Furthermore, dustbathing behaviour is not only regulated by peripheral factors such as feather lipid condition, but also by intrinsical factors (Vestergaard et al., 1990; Hogan et al., 1991; van Liere, 1991; van Liere and Wiepkema, 1992). In wood-shavings, rubbing may not be sufficiently reinforced, resulting in an enhanced expression of tossing (Fig. 1, Table 1). The correlation between the lipid condition and the amount of tossing in wood-shavings may therefore be a spin off from the underlying causality between rubbing a,ld tossing. The relation between the lipid condition and the amount of tossing may have been reinforced in the course of the experience, as bathing in this litter has some, be it marginal, effect (e.g. reduction of lipids of pennaceous feather parts: Fig. 2). Several factors may explain why not all feather lipid measurements after the dust deprivation correlated with the dustbathing quantities. Lipid quantity might have ambiguously interacted with lipid quality and/or with other related peripheral factors (e.g. fluffiness, skin irritation) in the different parts of the plumage. Moreover, the lipid level of the pennaceous feather parts of the breast increased due to the deprivation, whereas the lipid level of the pennaceous feather parts of the back did not increase. This was found for both substrate treatments (Fig. 2) and corresponds to lshida et al.‘s (1973) and van Liere et al.‘s (1991) findings. lshida et al. suggested that hens oil distal feather parts more than proximal feather parts. Van Liere et al. showed that hens oil the breast most of all regions of the plumage. Hens may therefore not equally appreciate lipid conditions of the plumage at different locations, The correlations between the lipid level of the one feather part and the lipid level of the other feather part were ambiguous, because they differed between the regions of the plumage, the times of feather sampling and the types of irtter. Therefore, detailed observations of oiling allocations are additionally necessary in order to fully understand the maintenance of proximal and distal feather lipids.
The valuable comments on earlier drafts of this paper by Professor Piet Wiepkema, the assistance of Thijs Roest, the helpful histological support by John Dulos and Professor L.P.M. Timmermans are greatly appreciated. The same is true for Kees Zoeter’s instructions concerning the English language.
eferences Anonymous, 1983. Animal feeding stuffs: Determination of diethyl ether extract. International Standard 5986, International Standard Organization, Geneva, 5 pp. Borchelt, P.L., 1975. The organization of dustbathing components in Bobwhite virginianus). Behaviour, 53: 217-237. Borchelt, P.L., Eyer, J. and McHenry Jr, D.S., 1973. Dustbathing in Bobwhite virginianus) as a function of dust deprivation. Behav. Biol., 8: 109-l 14. Borchelt, P.L. and Duncan, L., 1974. Dustbathing and feather lipid in Bobwhite virginianus). Condor, 76: 471-472. Borchelt, P.L., Hoffman, T., Hurrell, R.M. and McCarthy, R., 1979. Regulation of Japanese Quail (Coturnix coturnixh J. Comp. Physiol. Psychol., 93: 134-l 39.
quail (Colinus quail (Cohnus quail (Colinus dustbathing in
188 Conover, w.J., 1980. Practical Nonparametric Statistics. John Wiley and Sons, New York, 493 pp. Dow, D.D., I 988. Dusting and sunning by Australian brush-turkeys. Emu, 88: 47-48. Falsch, D.W., 1981. Das Verhalten von Legehennen in unterschiedlichen Haltungssystemen unter Berucksichtigung der Auf zuchtmethoden. In: D.W. Fijlsch and K. Vestcrgaard (Editors), Das Verhalten von Hiihnern, Birkhzuser Verlag, Basel, pp. 9-106. Healy, W.M. and Thomas, J.W., 1973. Effects of dusting on plumage of lapanese quail. Wilson Bull., 85: 442-448. Hodges, R.D., 1974. The histology of the fowl. Academic Press, London, pp. 648. Hogan, J.A., Honrado, G.I. and Vestergaard, K., 1991. Development of a behavior system: Dustbathing in the Burmese red junglefowl (Callus gal/us spacliceus): II. Internal factors. 1. Comp. Psychol., 105: 269-273. Ishida, K., Suzuki, T,, Kusuhara, S. and Yamaguchi, M., 1973. Influence of preen gland removal on the lipid over the plumage of roosters. Poultry Sci., 52: 83-87. Jacob, J. and Ziswiler, V., 1982. The uropygial gland. In: D.S. Farner, J.R. King and K.C. Parkes (Editors), Avian biology. vol. VI. Academic Press Inc., New York, pp. 199-324. Klinger, G., 1985. Untersuchungen zum Staub- oder Sandbade Verhalten von Hijhnern. lnstitut fiir Tierhaltung und Tierztichtting der Universit;it Hohenheim, PhD-thesis, pp. 119. Kruijt, J.P., 1964. Ontogeny of social behaviour in Burmese red jungle fowl (Callus &/us spadiceus). Behaviour Suppl., 12: l-201. Levine, R.L., Hunter, I.E. and Borchelt, P.L., 1974. Dustbathing as a regulatory mechanism. Bull. Math. Biol., 36: 545-553. Liere, D.W. van, 1991. Function and organization of dustbathing in laying hens. Agricultural University, Wageningen, PhD-thesis, pp. 123. Liere, D.W. van and Bokma, S., ‘i987. Short-term feather maintenance as a function of dust-bathing in laying hens. Appl. Anim. Behav. Sci., 18: 197-204. Liere, D.W. van, Kooijman, J. and Wiepkema, P.R., 1990. Dustbathing behaviour of laying hens as related to quality of dustbathing material. Appl. Anim. Behav. Sci., 26: 127- 141. Liere, D.W. van, Aggrey, S.E., Brouns, F.M.R. and Wiepkema, P.R., 1991. Oiling behaviour and the effect of lipids on dustbathing behaviour in laying hens Callus gal/us domesficus. Behav. Proc., 24: 71-81. Liere, D.W. van and Siard, N., 1991. Towards an understanding of litter bathing quality in hens. In: MC. Appleby, R.I. Horrell, J.C. Petherick and SM. Rutter (Editors), Applied animal behaviour: past, present and future. UFAW, Potters Bar, Herts, pp. 132-133. Liere, D.W. van and Wiepkema P.R., 1992. Effects of long term deprivation of sand on dustbathing behaviour in laying hens. Anim. Behav., 43. Lucas, A.M. and Stettenheim, D.R., 1972a. Avian Anatomy, Part I. Agricultural Handbook 362, U.S. Government Printing Office, Washington, pp. 750. Lucas, A.M. and Stettenheim, D.R., 1972b. Avian Anatomy, Part II. Agricultural Handbook 362, U.S. Government Printing Office, Washington, pp. 340. Norgaard-Nielsen, G. and Vestergaard, K., 1981. Dustbathing behaviour of uropygial gland extirpated domestic hens. Effects of dust deprivation. Acta Vet. Stand., 22: 118-I 28. Simmons, K.E.L., 1964. Feather maintenance. In: A.L. Thompson (Editor), A New Dictionary of Birds, McGraw-Hill, New York, pp. 278-286. Vestergaard K., 1981. Behavioural and physiological studies of hens on wire floors and in deep litter pens. In: D.W. F&h and K. Vestergaard (Editors), Das Verhalten von Hiihnern, Birkh;iuser Verlag, Basel, pp. 115-l 32. V-tergaad
6 1982. Dust-bathing in the domestic fowl; diurnal rhythm and dust-deprivation. Anim. Ethol., 8: 487-495.
Vestergaard,
K., Hogan,
J.A. and Kruijt, J.P., 1990.
Appl.
The development of a behavior system: red junglefowl. I. The influence of the rearing environment on the organization of dustbathing. Behaviour, 1 I 2: 99-I I 6. Zeman, P., 1988. Surface skin lipids of birds- a proper host kariomone and feeding inducer in the Poultry red mite (Demanysws galhe). Exp. Appl. Acarol., 5: 163-l 73. Dustbathing
in the hmnese
