DUSTBATHIN'G IN BOBWHITE QUAIL (COLINUS VlRGlNlANUS)‘: A REGULATORY MODEL Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY PETER L. BORCHELT 1972 ........... LIB'IX‘“ v Michigan 3 a to . Umvcrsity This is to certify that the thesis entitled DUSTBATHING IN BOBWHITE QUAIL (COLINUS VIRGINIANUS): A REGULATORY MODEL presented by PETER LEE BORCHE LT has been accepted towards fulfillment of the requirements for _2H.rD..——degree in Psychology I QMZXVCL “fig-MA ' Major professor 0-7639 ‘ amomc av § HMS 8r SUNS' BUUK BINULRY INE. . LIBRARY BINDERS I V I. seamen". meme»: [I - ea. AMI _ a, _ {W’ assoc of a\ avail been the c and 1 atic featl ance ABSTRACT DUSTBATHING IN BOBWHITE QUAIL (COLINUS VIRGINIANUS): A REGULATORY MODEL BY Peter L. Borchelt Dustbathing is one of a variety of biologically adaptive behaviors associated with care of the body surface which occur in a large number of avian species. Descriptions of dustbathing in many species are available, but few experimental investigations of this behavior have been reported. Borchelt, Eyer and McHenry (in press) briefly described the components of dustbathing in Bobwhite quail (Colinus virginianus) and reported that the frequency of some of the components showed system- atic increases with deprivation of dust. It was also observed that the feathers of birds which were deprived of dust had a more "oily" appear- ance than those of birds which had just dustbathed. A lipid regulation model was proposed for the function of dustbathing which stated that lipids from the uropygial gland were deposited on the feathers through "oiling" behavior to insure adequate lubrication of the feathers for maintenance of body temperature and for flight (Simmons, 1964). When the amount of lipids exceeds a critical level, the bird dustbathes. Duatbathing serves to remove lipids by driving dust into the plumage, after which the dust and lipids are vigorously shaken out of the plumage. The present experiments were designed to 1) describe in detail the organization of components of the dustbathing sequence in Bobwhite quail, and 2) test the proposed lipid regulation model by determining whether the amount of lipids on the feathers changes with deprivation of dust and whether surgical removal of the urOpygial gland leads to a decrease in frequency of dustbathing. ] dustbe pairs) 18 of of de; compon In dep T and th reliab interc coupon Signif with d Vation exbibi reveal nents, °f the of beh t°ry c. 11 °‘ 91. the am Standal lipids Borchelt In experiment 1, the frequencies of all of the components of the dustbathing sequence of each of 26 Bobwhite quail (housed in male-female pairs) were recorded, twice at 1 day of deprivation of dust, and for 18 of the birds, once at 5 days of deprivation. The two tests at 1 day of deprivation provided a measure of reliability of frequencies of components and the test at 5 days assessed the effects of an increase in deprivation on the frequencies of all components. The results showed that the frequencies of some of the components, and the sequence of first occurrences of the components, were highly reliable. The frequencies of some of the components were also highly intercorrelated, and the probable sequence of the first occurrence of components revealed a high degree of stereotyping for some components. Significant changes occurred in the frequency of many of the components with deprivation of dust, and two of the components showed sex X depri- vation interactions. The conditional.probabilities of some components exhibited time trends, and considering all occurrences of each component revealed a high degree of variability between birds in order of compo- nents. These results are discussed in relation to the classic definition of the fixed, or model action pattern, and in terms of the analysis of behavioral sequences into appetitive, consummatory and post-consumma- tory components. In experiment 2, 4 groups of Bobwhite quail were housed in groups of 9-13 and deprived of dust for either 1, 5, 15 or 180 days. After the appropriate deprivation period, the birds were sacrificed and a standard ether extraction procedure was used to assess the amount of lipids on a 2-3 gm. sample of the feathers of each bird in each group. The TI lipids days, suppo: of Ii] IEEOVI incis: group. the b: at we: inex In br three The pI tion 1 °f th. Purt‘m lead 1 suffi< 0f beI “We: Borchelt The results showed a significant increase from approximately 5 mg. of lipids per gm. of feathers at 1 day of deprivation, to 15 mg. at 5 days, to 35 mg. at 180 days of deprivation. These results strongly support the lipid regulation model which predicts a change in amount of lipids on the feathers with deprivation of dust. In experiment 3, 7 pairs of birds were divided into three groups. An experimental group (3 pairs) had the uropygial gland surgically removed while a sham-operated control group (2 pairs) recieved an incision on the back. Two pairs of birds formed an untouched control group. Starting one week after the experimental manipulations, the birds in all groups were given three tests at 1 day of deprivation at weekly intervals. These birds second test at 1 day of deprivation in experiment 1 served as a baseline to assess post-operative changes. In brief, no systematic changes between the baseline and any of the three post-operative tests were evident in any of the three groups. The possible addition of an "experience" factor to the lipid regula- tion model is discussed. These three eXperiments begin experimental analysis of a.care of the body surface behavior which occurs in a wide range of species. Further research investigating variables affecting dustbathing will lead to refinement of the lipid regulation model and may provide sufficient information to explicate the relations between the variety of behaviors associated with care of the body surface and yield a comparative analysis of this biologically important class of behavior. Date: . ‘0 g) L p Signed: K. ’49“, g $ng 2 /Q%uk MPM’. (Pr. DUSTBATHING IN BOBWHITE QUAIL (COLINUS VIRGINIANUS): A REGULATORY MODEL BY Peter LficBorchelt A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology To my wife ii for Dr. out let Dr. Com Lab Pou USE ACKNOWLEDGEMENTS My sincere thanks to Dr. Ralph Levine and Dr. Stanley C. Ratner for their advice and support as co-chairman of the committee. And to Dr. M. Ray Denny, Dr. L.I. O'Kelly and Dr. Robert K. Ringer for serving as members of the committee. Special thanks to Larry Duncan for assisting so admirably through- out all of the experiments. Thanks also to Stephen Overmann and Jerry Eyer for assistance in experiment 1. Special thanks also to Dr. Duane Ullrey, Mrs. Rosemary Covert, Dr. Donald Polin and Mr. Modestus Gomez for advice and assistance for work completed in the Nutrition Laboratory, Department of Animal Husbandry and the Department of Poultry Science, to Cynthia Haas for developing the computer programs used in experiment 1, to Dr. Robert K. Ringer for his helpful advice an experiment 3, and to Dr. Ralph Levine for his valuable assistance on the appendix. iii List List Genet Expe Gene Expe EXpe Gene Refe Apps TABLE OF CONTENTS List of Tables ........... ... ..................................... V List of Figures ....................................... ......... .. VI General Introduction ............................................. 1 Experiment 1 ..................................................... 8 General Discussion .......... ..... . ..... ..................... ..... 38 Experiment 2 ................... ..... ............................. 44 Experiment 3 ..................................................... 51 General Discussion ............................................... 57 References ............................................ ........... 60 Appendix ............... ..... ....... .......... ....... ............. 63 iv Table LIST OF TABLES Means, standard deviations and ranges of frequencies of test 1 components, and corre- lation coefficients (Pearson product moment) between frequencies of components of test 1 and test 2 (N I 24). Intercorrelations (Pearson product moment) among the frequencies of components. The average probabilities of first occurrences of components in relation to other components for test 1. The average probabilities of first occurrences of components in relation to other components for test 2. The average probabilities of first occurrences of components in relation to other components for test 3. A comparison, between subsequences, of the median conditional probabilities (f quartile deviation (Q) ) associated with the occurrences of components. A comparison of predictability of models I and II. A time course of values of FN’ 0N, and DN for models I and II. Page 15 16 21 22 23 26 74 75 Figure 1 2 l | Figure 10 11 LIST OF FIGURES An illustration of the lipid regulation model. An illustration of some components of dustbathing in Bobwhite quail. Mean frequency (1 standard error) of dustbathing components at one (test 2) and five (test 3) days of deprivation. The sequence of components and conditional probabilities associated with transitions between components (male number 1, test 1). The sequence of components and conditional prob- abilities associated with transitions between components (female number 1, test 1). The sequence of components and conditional prob- abilities associated with transitions between components (male number 2, test 1). The sequence of components and conditional prob- abilities associated with transitions between components (female number 8, test 2). The sequence of components and conditional prob- abilities associated with transitions between components (male number 4, test 1). Mean amount (+ standard deviation) of lipids on Page 12 18 28 30 32 34 36 the feathers 8f birds deprived of dust for 1, 5, 15, or 180 days. Mean number (+ standard error) of dust tosses, head rubs, and side rubs during the baseline and the three post-operative tests. An illustration of the original lipid regulation model. vi 46 54 64 -”“';jI| L: 11-3-14 . m In dust or 1 feet into Wing: Shap; relax and o the e GENERAL INTRODUCTION Although dustbathing is a behavior that occurs in a wide variety of avian species, it has received relatively little experimental investi- gation. General references to care of the body surface in birds (e.g. Simmons, 1964; Goodwin, 1956) mention dustbathing and describe the general sequence of behaviors involved. Simmons (1964) described this general sequence as follows: "Most dusting species form hollows of dust, if conditions allow, by squatting or lying down and performing movements of the bill (flicking, pecking), feet (scraping), and body (shuffling, rotating). The dust is driven into the plumage, either directly or indirectly, by movements of the wings (flicking, shuffling, shaking), or feet (scratching as in nest shaping or with one foot only), or of both wings and feet, the bird relaxing or ruffling its contour feathers, especially those of the rump, and often rubbing the head and bill in the dust. After dusting prOper, the earth is shaken out of the plumage, often vigorously." The usual medium for dustbathing is fine, dry earth, sand, or possibly dry rotten wood, although one report mentions observations of House Sparrows dustbathing in sugar (Goodwin, 1963). Speculations concerning the evolutionary origins of dustbathing have attempted to relate dustbathing to other care of the body surface behaviors. Chisholm (1944) postulated that dustbathing is the behavior from which anting was derived. Anting consists of movements whereby the to dis beh sta bat shi Hot We bet‘ has net] of l 0ft I’tlso 2 the defensive and other body fluids of ants (Formicidae) are applied t0' the birds feathers (Simmons, 1964). This explanation was discounted by Simmons (1957) on the basis that anting and dustbathing behaviors differ in form. Another speculation, by Nicolai (1962) states that dustbathing probably evolved from.water bathing. Dust- bathing has also been used in attempts to clarify phylogenetic relation- ships, for instance, between species of Columbidae by Nicolai, and Motacillidae by Master (1969), but without great success. Before phylogenetic comparisons can be fruitfully attempted and relations between care of the body surface behaviors discovered, research first has to be conducted focusing on the earlier stages of the comparative method (Denny and Ratner, 1970), which include precise description of behaviors, identification of productive preparations and explication of variables affecting the behavior of interest. A few studies have investigated variables affecting dustbathing in birds. Benson and Schein (1965) found that particulate surfaces (sawdust, soil, sand) elicited more dustbathing than solid surfaces (glass) in Coturnix, but the color of the substrate did not influence the incidence of dusting. Variations in air temperature and relative humidity did not significantly affect the incidence of dustbathing. Also, birds exhibited a ”satiation" effect, that is, birds maintained on a sawdust floor dustbathed in sawdust less than birds maintained on a wire floor. The effect of age of the bird on dustbathing has been investigated in Coturnix chicks (Brett & Kruse, 1967), White Rock chicks (Nice, 1962), and Burmese Red Jungle fowl chicks (Kruijt, 1964); these studies 89 Bo: Bot age chi des qua: soar incl squa onto side 3 generally show a greater incidence of dustbathing with increasing age. Borchelt (1971) reported that the development of dustbathing in Bobwhite (Colinus virginianus) chicks is influenced not only by the age of the chick, but also by the type of dust it encounters and the chick's experience with dust. A recent study by Borchelt, Eyer and McHenry (in press) briefly described the sequence of components of dustbathing in adult Bobwhite quail and found that deprivation of dust affected the frequencies of some of the components. The components of the dustbathing sequence include preliminary pecking and scratching movements in the dust, squatting in the dust, movements of the wings and feet to toss dust onto the birds' ruffled plumage (dust toss), rubbing the head and side in the dust (head rub and side rub), and ruffling of the feathers and shaking the dust out of the plumage (ruffle-shake). The sequence of components generally occurred in this order, although the precise order of the various components was quite variable. One aspect of the sequence was found to be highly stereotyped. After initial pecking, scratching, and squatting components, one or more dust tosses always occurred before the first head rub. In turn, one or more head rubs always occurred before the first side rub. No changes in the sequence of components were observed with an increase in deprivation of dust from 1 to 5 days. The frequencies of some components, however, increased significantly with an increase in deprivation level. The frequency of dust tosses and head rubs increased from 1 day to 5 days of deprivation, while the increase in frequency of side rubs, the decrease in latency to enter the SUCK EECE fEEE nent beha that of d feed appe, (per: also feat} CEnt of he feat} the 9 this z, the dust, and the difference in median numbers of seconds between successive dust tosses approached statistical significance. No differ- ences were found in the frequencies of components between male and female birds. Deprivation, then, is an important variable affecting some compo— nents of dustbathing, as it is for components of other classes of behavior, such as eating, drinking, and sexual behavior. This suggests that dustbathing functions as a regulatory system, with deprivation of dust leading to changes somewhere within the system, consequent feedback, and resulting compensatory changes in behavior. Borchelt, et al. suggest that the mechanism underlying such a system for dust- bathing is regulation of lipid substances on the birds' plumage. They observed that birds deprived of dust for 5 days had a more "oily” appearance than birds which had just dustbathed. Healy and Thomas (personal communication) have observed that Japanese quail (Coturnix) also have an "oily” appearance when deprived of sawdust for dustbathing. There are two sources of lipid on the plumage of birds. The feathers themselves contain endogenous lipid (from one to three per cent by weight depending on the species), probably as a by product of keratinization, a process occurring during the development of the feather (Bolliger and Varga, 1960). In addition, lipid material from the preen or urOpygial gland is applied to the surface of the feathers. This ‘oiling” behavior insures adequate water proofing of the feathers, maintenance of insulation, reduction of wear and chances of breakage, and possibly provides a source of vitamin 0 (Simmons, 1964). If some species of birds regulate the amount of lipid material on the feathers, Figt 1 EC rate duri eith. lipiI behar Iiays. the j by ar the c 1.99101} bathi regul Drovi Condi “huh. dust, f011m 5 a first approximation of how it might be accomplished is shown in Figure l (which is a simple non-linear regulatory mechanism; see Appendix 1 for details). "Oiling" behavior is assumed to continue at a constant rate, but when the amount of lipid increases above some critical level, (u) the bird dustbathes. Dustbathing presumably removes lipid from the plumage when the lipid is absorbed by the dust during the dust toss, head rub, and side rub components, and is shaken out of the plumage during the ruffle-shake component. It is assumed that the bird can either discriminate (by some unknown sensory mechanism) the amount of lipid on the feathers, or it can monitor the amount of its "oiling" behavior to determine the amount of lipid on the feathers. This lipid regulation model can be tested in a number of direct ways. If lipid from the uropygial gland elicits dustbathing, then the increase in dustbathing with deprivation of dust should be explained by an increase in lipid on the feathers with deprivation. Also, if the only source of additional lipid, the urOpygial gland, is surgically removed, then the frequency of dustbathing behavior should decrease. Three studies were designed to describe more clearly the dust- bathing behavior of Bobwhite quail and to test the proposed lipid regulation model for the function of dustbathing. The first study will provide a detailed description of the dustbathing sequence under two conditions of deprivation of dust. The second study will determine whether changes in lipid on the feathers occurs with deprivation of dust, and the third study will investigate the effects on dustbathing following surgical removal of the uropygial gland. .Hmeoe soaumaswmu wfiawa mfiu mo coaumuumsaafi Geom:_ $ e mo_>_._.«uaev mo whee Am umouv o>aw was Am umeuv use us musesoesoo wswcuenumsv mo Auouuo caduceus Hy essences“ seoZII.m ouswam 19 e $08 ON. 0! 0m. 00. m>_mn_wo n _ n _ n _ m _ L - 1 _ 1 row .m .ne . .m .n Lo... . $55 A $4193.13". .e 50 amp. . u u u q 1 1 sf“ 4 W 0 12 400 1 O 1 no: new 13 0 tom e #00. Joe . 48. use mew mam ch: 38. 58 .8 some 1 1 1 fl 4 1 d 1 L n a N. .. 1 D 1 @— 1 h 4 1 m EON #0. ...m ...N . 555m 2:8 .E xome 5:218 SBSNOdSBH :IO BBBWON NVBW 20 but the frequencies of the peck while squatting (F - 22.07, df = l, 16, p < .01) and the dust toss (F - 32.46, df - 1, 16, p < .01) components increased significantly with deprivation. The frequency of both the head rub and side rub components increased over trials (F - 27.78, df - 1, 16, p < .01; F- 24.18, df - 1, 16, p < .01, respectively), but, in addition, both of these components showed a sex X deprivation inter- action, with males showing more of an increase than females (head rub, F - 4.64, df - 1, 16, p < .05; side rub, F I 4.36, df - 1, 16, p < .06). The frequency of the ruffle-shake component increased significantly with deprivation (F - 24.72, df - 1, 16, p < .01). The increase in the frequencies of both the exit and other component approached statistical significance, and the frequencies of the enter component at each deprivation level almost exactly mirrored the frequencies of the exit component. Sequence of Compgnents A measure of the reliability of the sequence of components was devised by listing for each bird the order of the first occurrence of each of the components. Each component had been assigned a number from 1 to 13. Components missing from a birds' sequence were assigned the mean value of the components listed and included at the end of the list for that bird. This procedure was followed for the sequence of each bird at each test and yielded a very conservative measure of the test-retest reliability of the order of occurrence of components for each bird. The correlation coefficients thus computed between tests 1 and 2 ranged from .36 to 1.0 with a mean of .84. Another representation of the structure of the sequence of compo- nents is depicted in Tables 3, 4 and 5. These tables show the average 21 cm. mm. om. om. om. om. om. mm. mm. o.H ma om. oh. o.H mm. Hm. o.H om. o.H o.H no. o.H o.H Na on. No. mm. mm. mm. mm. Na. Na. om. ma. Ha 0H. mo. on. an. an. mm. mm. o.H mm. mm. o.a OH «H. ma. mm. on. o.H o.H o.H o.H «a. o.H o.H o.H ea. o.H o.H mm. o.H o.H .u umou you mumoooaaou menus ou sowuoaou mm. mm. o.H GA. «0. mo. ma. co. AN. on. o.H mm. mm. o.H Ca. Ca. wo. wo. mm. 0 on. 0 cm. om. mm. mm. o.H o.H m w nouu< mm. o.H no. no. no. no. no. no. no. we. o.H ma condo NH mxmsmlmawusm HH OH uwxm swam pom spam ppm omen meow upon :uuouum xuom unavm 300m nououom nouam ca nusooomaoe mo nauseousouo uuuau mo moauuafinmaoua ammuo>o one m manna ououom 22 NN. mm. mm. om. Hm. mm. mm. mo. na. ma. o.H o.H NH Nm. mm. no. no. nu. ma. o.H HH 0H. no. mN. No. On. mm. mm. o.H ow. o.H o.H OH NH. mm. mm. o.H o.H o.H o.H o.H cm. o.H o.H 0H. mo. mm. on. c.H Hm. o.H o.H oa. o.H o.H .N umflu HON mUflUflOQn—OU H0580 OH GOfiUflHUH 0H. mo. mm. mN. mm. mm. o.H Hm. o.H o.H 0H. no. mN. mo. mN. mu. o.H Hm. @H. 0H. no. no. mN. nN. NH. 0 mN. o HN. o o.H o.H o.H o.H n c eeee< NH. no. MN. «H. OH. OH. mo. ao. mo. «H. mm. o.H NH eerdo NH oxmnmIOHmmsm HH OH ume omHm mom ova now one: once Dana neuouum 300m umsvm xoom :uuauum mouom 6H mucosoaaoe no noosouusouo umuHu we moHuHHHnenoue owouo>u o:H o oHan uneven 23 Hm. o.H o.H mm. «a. «a. so. o.H o.H mm. o.H o.H NH mH. on. o.H mm. o.H o.H em. o.H o.H mm. o.H o.H NH ea. em. Nu. mm. mm. mm. am. am. mm. em. o.H HH mm. mm. cm. o.H o.H mm. o.H o.H 0H c.H «a. o.H o.H no. o.H o.H 00. NH. NH. o.H «a. o.H o.H mm. o.H o.H .m menu new mucocooaoo Hosuo cu :OHumHou co. NH. NH. He. no. o.H mm. o.H co. 00. NH. 00. co. 00. mm. mm. o.H mm. o.H o.H o o o o so. so. 0 o o o o 0 mm. 0 mm. o I o o.H . mm. mm. o.H o.H o.H o.H m e pede< No. No. NH. no. no. no. no. no. no. no. on. o.H NH ue50 NH oxmnmleHmmsm HH ume OH omHm m pom ova m ppm vmmm m away umso e euueeom n xuom a unadm m seem N :uumuom H Housm oH muaoaoaaou mo mousouuaooo umuHu mo moHuHHHnmnoua owouo>m one m mHan enemom 24 probabilities (across all birds) associated with the first occurrence of a component preceding or following the first occurrence of any of the other components for all three tests. In general, these tables reveal that the ordering of the first occurrence of some of the compo- nents is highly structured. The most stereotyped aspect of the sequence involves the first occurrences of the dust toss, head rub and side rub components. In agreement with the observations of Borchelt, et al., (in press), at least one dust toss preceded the first head rub, and at least one head rub preceded the first side rub. Also the first occurrence of the ruffle-shake component was highly likely to occur only after the first dust toss had occurred. The ordering of the other components was, however, less stereotyped. After the bird entered the dust, it was about equally likely for either the initial scratch or peck component to occur. The ordering of the other components was more variable, although patterns of probable occurrences are evident. The ordering of the scratch and peck while squatting components was most variable between tests, but neither these differences, nor any of the other differences between tests, were statistically significant (McNemar test for the significance of changes, Siegel, 1956). Tables 3, 4 and 5 oversimplify the organization of the sequence of components by only considering the order of the first occurrence of each component. Many components occur successively and considerable recycling between components also occurs throughout the sequence for each bird. A conditional probability matrix constructed for each bird (or group) would adequately present the probabilities of occurrence of com compon out th follou signif A duri ties u condit arity, an ent T0 tes were d- t0 firs first : quenCeg dust tc The CO] were C( SEQUEUI test 1 0CCurre Sec0nd U‘Tests betneeh lot thes 25 of combinations of components, but only if the probabilities of various components exhibited stationarity; that is, remained constant through- out the sequence. For instance, if the probability of component B following component A during the beginning of the sequence is not significantly different than the probability of it following component A during the middle and end of the sequence, the conditional probabili- ties would exhibit stationarity or show a lack of time trends. If the conditional probabilities for all of the components exhibited station- arity, a conditional probability matrix for all of the components for an entire sequence would accurately reflect the ordering of components. To test for stationarity, the test 1 and test 2 sequences for each bird were divided into five subsequences (enter to first squat, first squat to first dust toss, first dust toss to first head rub, first head rub to first side rub, first side rub to end of sequence). These five subse- quences correspond to the only components in the sequence (enter, squat, dust toss, head rub, side rub) whose order of occurrence was invariant. The conditional probabilities associated with the occurrence of components were computed for each of these subsequences. A comparison between sub- sequences of the conditional probabilities of some of the components on test 1 is presented in Table 6, which shows the median probability of occurrence (:_quartile deviation (Q)) of one component given that a second component has occurred. Significant differences (Mann-Whitney U—Tests, Siegel, 1956) occurred in the median conditional probabilities between some of the subsequences, clearly indicating a lack of stationarity for these conditional probabilities. 26 q _ He.NuN Am.~HV e.em Ae.NHV on HI 4 mm.eu~ Ae.me e.H Am.ev a.mN H] a ~m.~u~ Aw.mv m.w Ae.eav mN wN.NuN e mm.mue Am.eav mflmm Ae25v Rube he.mv a.me Ae.wmv a.mm me.eue . at we Ammo o on o ee.Hun on o ee.eun Aemv on oneez fleeces oneez oneez oncez cam new seem pom ewes mmoa omen eeaem OD and ova On new mom: ou mmOH umoa ou umsvm ou Houcm ouaoooomnom .mucmaoaSoo mo moommuuouuo ecu :uH3 woumHoommm AAOV QOHumH>oo oHHuumsv “w onuHHHAmnoua HM¢0HuHocoo cmeoE man we .mooooovomnsm consume .comHummEoo < 0 pH a.m.H. pea seem neem ewe: mmOH umsa loom comm ppm vmmm Immoa umoa chHuumovmv xoomlxoom chchmumv xoomlxomm k SINHNOJNOO 27 Two alternatives are available to represent the complexity and variability of the sequence of all of the components. An average order of components (along with average conditional probabilities and some indication of variance) could be constructed for all subsequences over all birds tested at each deprivation level. Such an average dustbathing sequence would, however, be too complex to easily depict and would not give a very clear picture of the large number of ways in which the components can be ordered, even within the constraints imposed by the high degree of stereotypy shown by some orders of components. Instead, the actual sequences of components and conditional probabilities associated with transitions between components within each subsequence for five birds are presented in figures 4 through 9. These individual sequences were selected to display the variability within sequences seen when the order of each occurrence of all components is considered and contrast sharply with the high degree of stereotypy revealed by considering only the order of first occurrences of each component. 28 Figure 4.--The sequence of components and conditional probabilities associated with transitions between components (male #1, test 1). 29 ENTER K SCRAT®33 / SOUAT\ r/j9/’ as scuarEE)ea 57 PECK 50 ea RISE "0 -- sane h ousr Toss PECK L° PE€E>JZ \ ..o J sounr I'-° scaarcuee Lo RISE: '°‘ 5 guano aua:; -°‘ 26 .oa@r T083 oz 22 .06 £§§§>‘ o3 up no PECK o3 scaarcu LO .76 o sonar RISE Heaosus [25 be @KTRUFFLE'Sfl‘KE .75 LOJ Jletcx , 1 J5 J4 U0 53 oust“ross RUFFLE‘SHAKE .33 s§\'9 sxnr eho cornea 30 Figure 5.-—The sequence of components and conditional probabilities associated with transitions between components (female #1, test 1). 31 ENTER LO .33 .OOPECK‘ ‘SCRATCHfiO .50 LO 0% 35 PECK4—-—SOUAT .l7 SCRATCH—LODUST TOSS——'3£—o RISE -5 25 .201 1.57 I.O ~7°PECK'—" SCRATCH4——l'g-—SOUAT _05 .43 25 ———o .SOHEAD RUB. PECK.53 .33 n ay’ .25 \ 05 ousrross' .04 .33 SIDE RUB——.OU3T TOSS .50 . .Is 33 .25 HEAD nus 07o \ .I4 2.5 SCRATCHo—OT—PECK _ I.O C125 RISE 1.0 EXIT—JLRurrts-SHAKE 20 .20 .60 .50 .25 ,75 '7 5° .50 Peck—'—»500Ar ENTERO—OTHER 33 1* , .75 t ' A. i SCRATCH .40 f 32 Figure 6.--The sequence of components and conditional probabilities associated with transitions between components (male #2, test 1). 33 ENTER LO ...Q. “\Qm ...- /SOUAT soSCRATCH/ so I0 up PECK 20 ~————Rlss PECK Je . g .n PECK HEAD “nus‘ .Oflfi ’I.O/ “ON SOUAT RISE ...... 1.. .. EHrtRH—Ql—rxur RUFFLE' SHAKE 9'— I .0. DUST 703$ ‘ :TI .77 \C “0 LO ENTER: .50 .67 EXIT be OTHER ADUST T039.” SIDER:.‘\\\i\\\\::sTTOO:S / RISE 34 Figure 7.——The sequence of components and conditional probabilities associated with transitions between components (female #8, test 2). 35 :ENTER EXIT SOUAT .20 /.o EXIT—ORUFFLE'SHAKE LO [.0 .33/0PEC5:(,—_. SCRATCH ' 4PE/C>£9 1... L——0'rHI:s 00 '0 SCRATCH .w \.15 yapecx 'M’ 0037 7039 'g I .II I 0 san7CH RISE-—— .25 L0 L67 PECK sxnr RISE 300Ar L0 ENTER ‘0 5° {43’ T0\\ L0 as —300Ar OTHER s7 sxur _£H7£H /O OTHER e4 4 as eon Rus' J7 20 N}, 3 T g, u T088 RISE Lo SCRATCH lto L0 I0 :OTHER .33 36 Figure 8.-—The sequence of components and conditional probabilities associated with transitions between components (male #4, test 1). 37 ENTER LO scseATCH LO SOUAT 33‘//h 43 ,‘ F—oSCRATCHg .25 'PECK-53 J0 /// RI3E=* '09 .03 Lo '0 ousTross ECK.67 SCIIATCH 03 '34 37 @TTOSSL .3I 'HEAo Ruse—J I \J SCRATCH 25 03 30 .m .2 -‘° J‘ o PECK.33 1’ h 33 .34HsA0 Rusi 'SIDE RUB.M) ‘ .Is .03 .m ousT T0185 .03 RUFFLE- SHAKEe~ ' EXIT 33 130 \\\¥ OTHER 50 gsouAT GENERAL DISCUSSION In summary, the results showed that the frequencies of some of the components, and the sequence of first occurrences of the components, were highly reliable. The frequencies of some of the components, par- ticularly the dust toss, head rub, and side rub components, were also highly intercorrelated, and the probable sequence of the first occur- rences of components revealed a high degree of stereotypy for some components. Significant changes occurred in the frequencies of many of the components with deprivation of dust, and the head rub and side rub components exhibited sex x deprivation interactions, with the frequen- cies of components increasing more for males than females. Dividing the entire sequence of components into subsequences revealed that the conditional probabilities of some components did not exhibit station- arity. If the sequence of all occurrences of each component within subsequences was considered, rather than only the first occurrence of components, a high degree of variability in orders and conditional probabilities was revealed. Does the fact that aspects of the sequence are reliable, some components are intercorrelated, and at one level of analysis the sequence of some components is highly sterotyped mean that the dust- bathing sequence in Bobwhite quail is a Fixed Action Pattern (FAP)? The defining characteristic of a FAP (Hinde, 1970, pg. 20-21) is that "although it may consist of a quite complicated spatiotemporal pattern of muscular contractions, (it) cannot be split into successive 38 res; ties diff stin con; rela IlESS sequ at 1 one demo with exhil neat, dEprj depri "Ould at tI. nents Droba of th °f th dustbe Once 1 39 responses which depend on qualitatively different external stimuli (Lorenz, 1935, 1937; Tinbergen, 1942)." Additional diagnostic proper- ties of the PAP (Barlow, 1968) are that it has common causal factors different from those of other fixed action patterns; once released, the stimuli triggering the PAP no longer exercise control over it; and its components appear in a predictable sequence in time with the inter- relation of the parts constant, even though the "intensity" or complete- ness of the components may vary. At one level of analysis, some components of the dustbathing sequence do appear in a perfectly predictable temporal pattern; that is, at least one dust toss always preceeds the first head rub and at least one head rub always preceeds the first side rub. There is also evi— dence that the completeness of the sequence of these components varies with deprivation. At one day of deprivation (test 1) all (24) birds exhibited the dust toss component, 23 birds (962) the head rub compo- nent, and 20 birds (832) the side rub component. At five days of deprivation (test 3) all birds exhibited each of these components. At deprivation levels less than one day, it is probable that fewer birds would exhibit the complete sequence of these components. Of course, at the level of analysis of the first occurrence of all of the compo- nents, the sequence ceases to be perfectly predictable, and merely probable. At yet a further level, considering each occurrence of all of the components, it is not possible to accurately predict just which of the components will be the next to occur. It may be difficult to maintain that the eliciting stimuli for dustbathing cease to exert control over the expression of the behavior once it has been released. If one considers the dust as the eliciting stir seql whil con; shat this any anal then the I dept: same form 91th etty Nb Q tiOns compo 40 stimulus, evidence for it continuously modulating the dustbathing sequence may be found in the high intercorrelations between the pecking while squatting component and the dust toss, head rub, and side rub components. The pecking component presumably loosens, or perhaps even shapes, the dust so these components can occur. Of course, this refers to the dust modulating successive occurrences of these components in this manner; determining whether or not the dust modulates the form of any single instance of one of these components will require much finer analysis. If one accepts that lipids on the plumage elicit dustbathing, then successive occurrences of these components are again influenced by the eliciting stimulus, since a (hypothetical) increase in lipids with deprivation of dust elicits a greater frequency of components. The same uncertainty, however, exists concerning the effect of lipids on the form of a single occurrence of a component as exists for such on effect with dust. Some components meet the requirement of the last diagnostic pr0p- erty of the PAP (Barlow, 1968) since the dust toss, head rub, and side rub components are high intercorrelated, suggesting common causal func- tions. Assuming the lipid regulation model to be valid, these three components are the only ones which actually serve to get dust into the plumage. whether or not the dustbathing sequence fits the formal definition of the PAP, that it cannot be split into successive responses which depend on qualitatively different external stimuli (Hinde, 1970) depends upon which components are considered. They may meet the requirement since non-particulate stimuli (Benson and Schein, 1965; personal obser- vation) do not elicit these components and presumably neither do 41 substances on the feathers other than lipids from the uropygial gland, although this has not been systematically tested. Whether quantitative differences in the type of dust or chemical composition of lipids makes a difference in measures of these components awaits further research. The other components that occur during the sequence such as pecking and scratching 253_elicited by a number of qualitatively different stimuli, such as food or aggressive encounters. Thus, it appears that the dust toss, head rub, and side rub compo- nents fit the formal definition of a RAP and meet many of the properties described by Barlow (1968). However, the suggestion by Marler and Hamilton (1966) and Barlow (1968) that the term "Fixed Action Pattern" be replaced with Modal Action Pattern (MAP) is supported by the results of this study. The "fixedness" of the pattern of dustbathing components depends entirely on which level of analysis is chosen. Of course, in the present study, only the frequencies and sequences of these compo- nents were measured. Further research, perhaps using films of the sequence and recording the form, duration, and other measures of the components, will yield additional information concerning the degree of variability around this modal action pattern. The preceding analysis of the dustbathing components can be viewed from another framework, namely the subdivision of behavioral sequences into appetitive, consummatory, and post-consummatory components (Denny and Ratner, 1972). Appetitive components of behavioral sequences are movements associated with orientation to and selection of particular stimuli, and are relatively variable, both for an individual and between individuals of the same species. The consummatory components correspond to the fixed action pattern (Lorenz, 1950) and are more sterot rub, a of the retest freque two of coupon compon exhibi instan while genera frEque Signif increa C°mP0n Sests consun and su betwee c°mP0n diseng meat 1 T quency 42 sterotyped than the appetitive components. Thus, the dust toss, head rub, and side rub components can be considered consummatory components of the dustbathing sequence; these components exhibited high test- retest correlation coefficients and were highly intercorrelated, the frequencies increased with deprivation of dust, and the frequencies of two of the components showed sex X deprivation interactions. Although the distinction between the appetitive and consummatory components is generally difficult to make (Denny and Ratner, 1970), the components other than the dust toss, head rub, and side rub components exhibited different features from these consummatory components. For instance, although one of the initial components of the sequence, peck while standing, had a high test-retest reliability coefficient, it was generally not highly intercorrelated with other components, and the frequencies of both the initial peck and scratch components showed non- significant decreases with deprivation of dust in contrast to the increase in frequencies of other components. This difference between components affected and those not affected by deprivation of dust sug- gests that the initial peck and scratch components are appetitive to the consummatory dust toss, head rub, and side rub components. The squat and subsequent peck and scratch components are intermediate components between appetitive and consummatory components, and the ruffle-shake component is presumably a post-consummatory component which serves to disengage the bird from the dustbathing sequence preparatory to engage- ment in other consummatory behaviors such as eating and drinking. The finding that deprivation of dust led to increase in the fre- quency of most components of the sequence replicates and extends the research of Borchelt, et. al, (in press). Observations made during 43 the present study also indicated that there was an increase in lipid on the feathers when Bobwhite quail were deprived of dust for 5 days. These observations lend further support to the hypothesis that lipid on the plumage elicits dustbathing. The finding that there is a sex x deprivation interaction for the head and side rub components further supports this hypothesis. Kar (1947) reported that the secretion of the uropygial gland in domestic chickens was influenced by testosterone. This would suggest that if lipid fron the uropygial elicits dustbathing, then there should be a sex difference in some measure of dustbathing, with males showing more dustbathing than females. A direct test of the lipid regulation model is the purpose of the next study, which will quantify the observed increase in lipid on the plumage of Bobwhite quail deprived of dust. Ca EXPERIMENT 2 Introduction Experiment 1 replicated and extended the results of Borchelt, et al., (in press) by finding that the frequencies of many of the (‘1! components of the dustbathing sequence increased with an increase in the level of deprivation of dust, and that the frequencies of the head ( and side rub components increased more for males than females. The 1 purpose of the present study is to determine whether the mechanism for this deprivation effect is lipid substance from the urOpygial gland by quantifying the observed change in lipid on the plumage with deprivation of dust. Subjects and Procedure Sixteen male and twenty-six female Bobwhite quail (Colinus virginianus) approximately six months of age were divided into 4 groups. Group 1 (4 Of 5 Q) was deprived of dust for 1 day, group 2 (5 O: 6 Q) for 5 days, group 3 (2.61 7 Q) for 15 days, and group 4 (5 of 8 Q) was deprived for 180 days (never allowed access to dust). Prior to the appropriate level of deprivation, groups 1, 2 and 3 were given 7 days continuous access to dust. Groups 1, 2 and 3 were raised from chicks in the same laboratory as birds in experiment 1 and were housed in cages 137 x 50 x 30 cm. on a 14:10 light-dark cycle. Groups 4 was maintained at the Department of Poultry Science Farm under continuous light in a 100 x 75 x 24 cm. cage with approximately fifty other birds until 5 months of age when 44 45 they were transferred to a cage identical to the cages of the other groups. Food (King Milling Company, Lowell, Michigan, U.S.A.) and water were continuously available at all times to each group. After the appropriate period of deprivation of dust, all birds in each group were sacrificed (ether asphyxiation). The feathers of each bird were then cleaned with compressed air to remove any remaining dust from the plumage. The distal portion of most of the feathers (except the primaries) of each bird were clipped off with scissors, leaving the calamus intact. A 2-3 gm. sample of feathers from each bird was weighed and subjected to a standard ether extraction proce- dure on a Goldfisch apparatus for 2 hours. The collected lipid and ether were poured into tared containers, the ether evaporated, and the remaining lipid weighed on a Mettler analytical balance to an accuracy of 1 mg. Replicate samples of feathers were run from birds in groups 1, 2 and 4. No differences were found between replicates and the data are combined for each group. Results The results are shown in Figure 9 which depicts the mean amount (1 standard deviation) of lipid, expressed in mg. lipid per gm. feathers, for each of the groups. A one-way analysis of variance revealed the change in amount of feather lipid with deprivation of dust to be highly significant (F . 79.4, df - 3/38, p < .001). Comparisons between individual groups using the Newman-Keuls proce- dure (Winer, 1962) indicated that all of differences were highly significant (p < .001) except for the difference between groups 2 and 3, which was not significant. No difference was found in amount of lipid between male and female birds. 46 .mame owe so .mH .m .H pom umae mo vo>wuaou woman mo mumpumom oSu so mwfiafia mo Assaumw>wo vumcamum My unseen ammzlu.m madman SAVCI NI NOLLVNHdSO 9| S'N OQI EI'N 47 LIPIDS (mg/gm feathers) +01 N 0| 0 O ‘O‘b 6'N II'N 48 Discussion The results clearly show that between 1 and 5 days of deprivation of dust there was a significant increase in feather lipid, confirming the observations of Borchelt, et al., (in press) and Experiment 1, and providing strong support for a lipid regulation model of dust- bathing in Bobwhite quail. ?*D The results of the present study have to be considered in view I of reports that avian plumage contains lipid other than secretions from the uropygial gland. Bolliger and Varga (1960, 1961) and Bolliger and Gross (1958) have reported an average of approximately 2 percent a- 4-»... _ 4 . 1 4.; .x total feather lipid (percent of dry feather weight) in a number of species of birds. These investigators suggest that the endogenous lipid is probably formed as a by—product of keratinization. The results of the present study, in which percent of total feather lipid ranged from 0.5 to 3.5, indicates that such endogenous feather lipid is certainly supplemented by lipid from the uropygial gland and the total amount of lipid can be regulated by dustbathing. The lack of a significant increase in feather lipid between 5 and 15 days of deprivation suggests that there may be no increase in the frequency of dustbathing components between these two deprivation levels. This could occur if the frequency of "oiling" behavior levels off between 5 and 15 days of deprivation although the frequency of "oiling" does not reach an asymptote at 15 days since after 180 days of deprivation there was again a significant increase in feather lipid. It is also possible that this finding is due to a difference in con- version of dietary fats to lipid secretion from the uropygial gland 49 (Apandi and Edwards, 1964) since the weights of the birds in group 3 averaged about 202 less than in the other groups (160 gm. vrs. 200 gm.). Further research investigating both the changes in dustbathing behavior between 5 and 15 days of deprivation, and the relations between dietary fats, body weight, and feather lipid, will be necessary to resolve this discrepancy. The lack of a sex difference in amount of feather lipid is surprising in view of the sex differences found in the head and side rub components in the previous study. The procedure in the present study for depriving the birds of dust, however, was different than in experiment 1. Since the birds in this study were housed in large groups and since groups of Bobwhite quail form dominance hierarchies, the birds were given one week of access to dust prior to deprivation to reduce the chance that birds low on the hierarchy would not have an Opportunity to dustbathe. This is in contrast to the two days access to dust given in experiment 1. An alternative explanation for a lack of a sex difference is that feathers from the entire body surface of the birds were analyzed. It is possible that the smaller feathers of the head and flanks of the bird differed in amounts of lipid for males and females, but this difference was obscured by the larger percentage of feathers from the breast and back of each bird. Support for this view is found in data presented by Bolliger and Varga (1960) showing differences in the chemical composition of feather lipid between small and large feathers of an unspecified species of duck. Clearly, biochemical anlysis of the lipid from both the uropygial gland and plumage of one species of bird will be necessary to elucidate the relations between sex, uropygial and feather lipid and dustbathing. 50 These results also cast doubt on the generally accepted thesis that dustbathing serves primarily to remove ectoparasites. (Simmons, 1964; Goodwin, 1956; Stoddard, 1931). No ectoparasites were observed on the feathers of birds in this or in previous studies (Borchelt, et al., in press; Experiment 1). Since the principle food of ectoparasites is lipid substance from the feathers (Dubinin, cited in Kelso and Nice, 1963), dustbathing could perhaps secondarily remove ectoparasites by reducing their food supply, as well as by dessicating them or interfering with their respiration. It is clear that dustbathing occurs in Bobwhite quail which do not have ectOparasites and the lipid regulation model is offered to explain the function of dustbathing in Bobwhite quail, and perhaps other avian species as well. EXPERIMENT 3 Introduction The results of experiment 2 provide strong support for the lipid regulation model for the function of dustbathing in Bobwhite quail. The purpose of the present study is to further test the model by examining the effects on dustbathing following surgical removal of the main source of lipid on the plumage, the uropygial gland. Elder (1954) reviewed a number of studies which attempted to determine the function of the uropy- gial gland by surgically removing it. The general conclusion, supported by some of his own experiments on ducks, is that removal of the uropy- gial gland leads to gradual deterioration in the condition of the plum- age (which is temporarily improved with molting) resulting in faster wetting of the feathers, scaling and peeling of the skin and bill, and a lower rate of growth for glandless as compared to normal birds, pre- sumably due to a greater energy loss in glandless birds as a result of a less efficient insulating layer of feathers. Thus, the lack of lipid from the urOpygial gland adversely affects the condition of the plumage; the temporary improvement with molting would be due to a renewal of feathers containing endogenous feather lipid. No behavioral measures have been made on such Operated birds other than Elder's (1954) note that there was no reduction in "oiling" behavior when the uropygial is removed. According to the lipid regulation model, howb ever, removal of the urOpygial gland should lead to a decrease in dust- bathing behavior since there is a decrease in lipid applied to the feather. 51 52 Subjects: Seven pairs of Bobwhite quail used in experiment 1 were used as sub- jects for this experiment. Three pairs were randomly assigned to the experimental (operated) group, and two pairs each were randomly assigned to an undisturbed control group and a shamroperated group. Procedure: The eXperimental birds were individually removed from their cages, : placed in a holder, and a topical anesthetic (Cetacaine) was applied to i the skin surrounding the uropygial gland. As much of the uropygial gland g as possible was surgically removed, the skin sutured, and the birds given : a week to recover from the operation. Post-mortem examination of these :-u’ birds after the experiment revealed that there was no remaining lipid material from the uropygial gland. The birds in the sham-Operated con- trol group were treated identically, except the anesthetic was applied to a small area of skin on the back of the bird, a small cut was made and sutured, and the bird was returned to its cage. These birds were also given one week for recovery. The control birds were left undis- turbed in their cages during this time. The results of the second test at one day of deprivation (experi- ment 1) were used as the baseline data for all of the groups. One week after the apprOpriate operations were made, all of the birds were tested again at one day of deprivation with two days access to dust prior to deprivation. Two additional tests were given at weekly intervals for all groups. The recording procedures remained the same as in experi- ment 1 for all of the tests. 53 Results and Discussion Each bird engaged in at least one dustbathing sequence on each of the three post-operative tests. Only the dust toss, head rub, and side rub components of the dustbathing sequence were analyzed since these are the only components directly involved in getting dust onto the birds' plumage. Figure 10 presents the results of the baseline test and the three post-Operative tests for each of the three components. No statis- tical analyses were made on these results, since it is clear that there was no systematic decrease in the experimental group, and that the sham- operated group did not recover to baseline levels during any of the three post-Operative tests. The results of this study are disappointing in that they showed no systematic decrease in the frequency of either the dust toss, head rub, or side rub components in birds which had the uropygial gland removed. These birds, however, had extensive experience with dustbathing. It is possible that the lipid regulation model, as illustrated in Figure 1, may have to include an "experience" factor in the feedback loop between the uropygial gland and dustbathing (see Appendix). This possibility can be tested by removing the uropygial gland from different groups of birds either experienced or inexperienced with dustbathing. An additional factor may also be experience with "oiling"; this would necessitate removing the uropygial gland from different groups of birds either before or after experience with "oiling" ("Oiling" is first seen at about 17 days of age in Bobwhite; Nitschke, 1972). This explanation of the results of the present study is quite plausible since other instances of interactions be— tween experience and the behavioral effects of hormones occur in other classes of consummatory behaviors, for instance, sexual behavior (Rosenblatt .mummu o>fiumumm0Iumoa momma mnu mam mafiaommn onu weapon was» spam paw .mp2» poms .mommOu umsp mo Amouuo pumpamum Hv panama smmflll.oa ouswam 55 FREQUENCY MEAN D mxvmmzsmzdfir 0 0024.307 .00 Dm1>2-00mm>4m0 .00 mflom 1l0 3:0 mo 10 0 1U 40 RF Im>0 00 10.1 LA 200 u0 wO +vV/T OCWH N0 40mm 10 .0 mbmmsz N u ...mmd 56 and Aronson, 1958). A reasonable prediction from this explanation would be that in— creasing the deprivation level for the birds in this experiment would have resulted in no deprivation effect if they were dustbathing only due to extensive past experience rather than to remove lipid from the feathers. This also implies that there would have been no increase in lipid with deprivation of dust for these birds. Lipid could be experi- mentally applied to the birds' feathers to determine whether it could again elicit increases in dustbathing. ‘A'!.," - ‘ "elA-A 5.1x n-;;—A GENERAL DISCUSSION The results of these three studies suggest some interesting areas for future research. First of all, dustbathing is a very robust behavior, with all birds engaging in a dustbathing sequence even at low levels of deprivation. Dustbathing is a good model for analysis of the structure of behavioral sequences since it contains a large number of components, some of which, at least in terms of first occurrences, are relatively stereotyped. Observation during these experiments also suggests that the durations of some of the dustbathing components are relatively constant. Of course, analysis of films of these components would be necessary to determine the average duration and degree of variability of components. Benson (1965) filmed some of the dustbathing components of Coturnix quail at 200 frames per second and found that the duration of the dust tossing component averaged 1.12 seconds for one bird and 1.13 per another; unfortunately, no data on the variability of these durations are reported. Use of films would allow for a much finer analysis of individual differences as well as degree of stereotypy and variability. Such analyses are not only theoretically interesting for their information concerning either the validity of conventional definitions of fixed (or modal) action patterns, or for their use in analysis of appetitive, consummatory, and post- consummatory components of behavioral sequences, but they may also yield information helpful for making finer classifications of behavior. In addition, they may also serve as convenient baselines to assess the effects of variables on aspects of the behavioral sequence. The second experiment provided compelling support for the lipid 57 58 regulation model, but future research could provide more direct evidence for the model by experimentally manipulating the level of lipid on the plumage. Several commercial products suitable for this purpose are availa— ble. A vegetable oil product (Pam) in aerosol cans could be Sprayed in measured amounts onto the plumage, which should lead to an increase in dustbathing behavior. Several products are also available for removing lipid from human hair and would probably function the same for urOpygial lipid. While these are highly artificial means of modifying the amount of lipid on the feathers of birds, positive results would greatly strengthen support for the lipid regulation model. Suggestions can also be made concerning future areas of concentration for field studies of dustbathing and other COBS behaviors. If the lipid regulation model is valid, then differences in lipid production between male and female birds is expected. This suggests, of course, that differ- ences should occur between males and females in dustbathing and perhaps other COBS behaviors. Seasonal differences should also occur, since photo— periods strongly affect production of reproductive hormones (Sturkie, 1965). Field studies to date have generally not reported data concerning sex dif- ferences in COBS behaviors because such differences have not been looked for. The third study was less conclusive than the other two. Future research should consider the experience of the bird with both dustbathing and "oiling", and should assess the level of lipid on the feathers and re- cord dustbathing behavior at different deprivation levels for birds with the ur0pygia1 gland removed. An alternative method for manipulating the amount of lipid from the urOpygial gland is removal of the testes, which eliminates production of testesterone, which in turn should reduce the sex difference found in some measures of dustbathing. In 59 In summary, these studies begin behavioral analysis, using the stages of the comparative method (Denny and Ratner, 1970), Of a system or class of behaviors which have previously only been described. Guiding questions for future research should focus on determination of the mech- anisms underlying the Operation Of this behavioral system with the goal of discovering whether the system functions on the same principles of Operation as do other systems. Such analyses on these and other behavioral systems, on a variety of species, are necessary to attain a general, comparative psychology. REFERENCES REFERENCES Altmann, S. A. Sociobiology of rhesus monkeys, II. Stochastics of social communication. J. Theoret. Biol., 1965, 8, 490-522. Apandi, M. and Edwards, H. M. Studies on the composition of the secretion Of the uropygial gland of some avian species. Poultry Science, 1964, 43, 1445-1462. Barlow, G. W. Ethological units of behavior. In D. Ingle (Ed.) The Central Nervous System and Fish Behavior. University Of Chicago Press, Chicago, 1968. Barlow, G. W. Ethology of the Asian Teleost, Badis badis. IV. Sexual behavior. COEeia, 1962, 2, 346-360. Benson, B. N. Dustbathing in Coturnix quail (Coturnix coturnix japonica). Unpublished Master's Thesis, The Pennsylvania State University, 1965. Benson, B. N. and Schein, M. W. Factors affecting dust-bathing in Coturnix quail. Amer. ZOOl. (Abstr.), 1965, 5(2), 146. Bolliger, A. and Varga, D. Feather lipids. Nature, 1961, 4781, 1125. Bolliger, A. and Varga, D. Cholestanol in avian plumage. Austral. J. 2&0 3101.. 1960, E, 265—2700 Bolliger, A. and Gross, R. Ether soluble compounds associated with the plumage of birds: A sterOid constituent. Austral. J. exp. Biol., 1958, 36, 333-338. Borchelt, P. L. DevelOpment of dustbathing in Bobwhite (Colinus virginianus). Paper presented at 89th annual meeting Of American Ornithologist s Union, Seattle, Washington, 1971. Borchelt, P. L., Eyer, J., and McHenry, D. S. Dustbathing in Bobwhite quail (Colinus virginianus) as a function of dust deprivation. Behavioral Biology, in press. Brett, W. J. and Kruse, M. DevelOpment Of dusting pattern in young quail (Coturnix coturnix japonica) Proc. Ind. Acad. Sci., 1967, 15, 282. Chisholm, A. M. The problem of anting. Ibis, 1944, 86, 389—405. Delius, J. D. A stochastic analysis of the maintenance behavior of skylarks. Behaviour, 1968, 137-177. Denny, M. R. and Ratner, S. C. Comparative Psychology, revised edition. Dorsey Press, Homewood, 1970. 60 A out r—nm-— .~__—.____ ... .-.u._-_ .9 61 Eisenberg, J. F. The behavior of Heteromyid rodents. Univ. California Publ. 2001., 1963, 62, 1-100. Elder, W. H. The oil gland of birds. Wilson Bull., 1954, 66, 6-31. Goodwin, C. House sparrows dustbathing in sugar. Brit. Birds, 1963, 66, 378-379. Goodwin, D. Care of the body surface--preening, bathing, dusting, and anting. In H. P. W. Hutson (Ed.), Ornithologists Guide. Brit. Ornith. Union, London, 1956. Hinde, R. A. A comparative study of the courtship of certain finches (Fringillidae). Ibis, 1955/56, 21, 706-745; 22, 1-23. Hinde, R. A. Animal Behaviour, Second edition. McGraw-Hill, New York, 1970. Rat, A. B. The hormonal influence in the normal functioning of the uropygial gland in the fowl. Anat. Rec., 1947, 22, 75-89. Kelso, L. and Nice, M. M. A Russian contribution to anting and feathermites. Wilson Bull., 1963, 16(1), 23-26. Kruijt, J. P. Ontogeny of social behavior in Burmese Red Jungle-fowl. Behaviour Supplement, 12, 1964. Lorenz, K. Der Kumpan in der Umwelt des Vogels. J. fur Ornith., 1935, 62, 137-213, 289-413. Lorenz, K. Uber die Bildung des Instinktbegriffes. Naturwiss, 1937, 26, 289-300, 307-318, 324-331. Lorenz, K. The comparative method in studying innate behavior patterns. In Symp. Soc. exp. Biol. IV, Physiological Mechanisms in Animal Behaviour, 1950, Cambridge University Press, Cambridge, England. Marler, P. and Hamilton, W. J. Mechanisms of Animal Behavior, Wiley, New York, 1966. Mester, H. Staubaden und Einemsen bei Piepern. J. fur Ornith., 1969, 110, 487-492. . Nice, M. M. DevelOpment Of behavior in precocial birds. Trans. Linn. SOC. NOYO’ 1962’ g, 1.211. Nicolai, J. Uber Regen-Sonnen-und Staubaden bei Tauben (Columbidae). J. fur Ornith., 1962, 103, 125-139. Rosenblatt, J. S. and Aronson, L. R. The influence of experience on the behavioral effects of androgen in prepuberally castrated male cats. 62 Siegel, J. Nonparametric statistics for the behavioral sciences. McGraw-Hill, New York, 1956. Simmons, K. E. L. Feather maintenance. In A. L. Thompson (Ed.) 6 new dictionary Of birds. McGraw-Hill, New York, 1964. Stoddard, H. L. The Bobwhite quail: its habits, preservation, and increase. Charles Scribner's Sons, New York, 1931. Sturkie, P. D. Avian Physiology, Second edition. Comstock, Ithaca, 1965. Tinbergen, N. An Objectivistic study of the innate behavior of animals. Biblioth. Biother., 1942, 1, 39-98. Winer, B. J. Statistical Principles in Experimental Design. McGraw-Hill, New York, 1962. APPENDIX APPENDIX The original formulation of the lipid regulation model is shown in Figure 11. This model assumes that both dustbathing and oiling be- haviors are regulated by the bird to maintain within certain limits the amount Of lipids on the plumage. Observations during both the Borchelt, Eyer and McHenry study and experiment 1 indicated that oiling behavior is not regulated since it does not decrease with deprivation of dust. These Observations thus conflict with the original lipid regulation model. However, by constructing a mathematical formulation of this model, it can be discovered that the graphic representation (Fig. 11) conflicts with the results of experiments 1, 2 and 3 as well. The first mathematical model presented in this section illustrates this point. 1k: contrast, a second mathematical model was constructed to correspond to the lipid regulation model illustrated in Figure l, which assumes that dustbathing is regulated while oiling is maintained at a constant rate. Some implications of this second model will be discussed. Original Formulation Model I was formulated as follows: Let: FN = Amount Of lipids on feathers at time N ON = Amount of oiling behavior at time N DN = Amount of dustbathing at time N u = Critical level for amount Of lipids on feathers 63 l"““““i‘“+f7. I 5 - - ,. . In 64 .Hmwoe cowumaswou oHeHH Hmcwwfiuo ozu mo sowumuumsaafi Oaoma p e Imo_><1mm oz_s:i_oz_:e 0) Fn+1 - ON+1 I 0, if FN > u “(U'FN), if FN < u DN+l I 0, if FN < u Suppose the parameters for these equations are: O I 2 (II B I 1 u I 5. This model yields the following time course with starting values of F I 10, O I 0, D I 0. The time units are N N given in days. Time FN ON DN 0 10 O O 1 10 O 10 2 0 0 10 3 0 10 0 4 10 10 0 5 20 O 10 At this point, dust is removed and the birds given one day of deprivation of dust. 6 10 O 10 Now the birds are given access to dust. Test 7 10 0 10 Using the same initial values prior to deprivation the birds in another experimental group are given five days of depriva- tion of dust and then access to dust. 67 Time FN ON DN 0 20 O 10 Five 10 0 0 days 10 0 O Of 10 O O dust 10 0 0 deprivation lO 0 0 Test 6 10 0 10 As can be seen, the value of DN is 10 units after both one and five days of deprivation, which conflicts with the results of the Borchelt, et a1. study and experiment 1. It is also obvious that the value of FN does not increase with deprivation of dust, which conflicts with the results of experiment 2. This model predicts the following for experiment 3: Time FN 0N DN baseline Test 0 20 0 10 Remove urOpygial gland one week of 10 O 0 recovery (depri- 10 0 0 vation) 10 0 0 Access to dust 8 10 0 10 Access to dust 9 0 O 10 10 0 O 0 Post-Operative Test-1 ll 0 0 0 68 These values continue for all subsequent post-operative tests. The data from experiment 3 do not correspond to the predictions of Model 1. Model I predicts a decline in the value of D to zero N on the first and subsequent post-Operative tests. Alternative Starting Values Starting with different values will indicate the generality Of the predictions of Model I. With initial values of FN I 0, O I O, N DN I 10, it can be seen what the model will predict in the case where the initial level Of lipids on the feathers is below the critical level, n. Time FN 0N DN 0 0 0 10 At this point dust is removed and the birds given one day of deprivation of dust. l 0 10 0 Now the birds are given access to dust. 2 10 10 O In another group of birds. 0 0 O 10 Five 0 10 0 days 10 10 O of 20 O 0 dust 20 O O deprivation 20 0 0 Test 6 20 0 3O 7 -" u .35.“... 7‘ n- emawnrzs-vnnJ-u .. 69 If the initial value of FN is below u, then there will be an increase in both FN and DN with an increase in deprivation of dust. This increase will occur, however, at any time after three days of deprivation and the values of FN and DN will not continue to increase. Thus, with these starting values, Model I does not correspond to the results of the Borchelt, Eyer and McHenry study which found an increase in frequency of dustbathing from 1 to 3 days Of dust deprivation, and a continued increase to 5 days of deprivation. Model I also does not predict, with these initial values, the results of experiment 2 which showed continued increases in the value of F . Unfortunately, N Model I predicts an eventual leveling off of the values of FN and DN with time. What about experiment 3? Time FN ON DN baseline 0 0 0 10 Test remove Oil gland one week Of 0 O 0 recovery (depri- 0 0 0 vation) 0 O 0 Access to dust 8 0 0 0 Access to dust 9 0 O O Deprivation 10 O O 0 Post-Operative Test-1 11 0 O 0 fl Tn.‘ 1—.'-.wu— 70 These values continue for all subsequent post Operative tests. With these new initial values, Model I still does not predict the results Of experiment 3. Alternative Formulation From these considerations, as well as the observations that fre- quency of oiling behavior does not decrease with deprivation of dust, it is clear that a new model is called for. Model II incorporates a constant level of oiling behavior. A second constant sets a minimum level of DN due to functions other than lipid regulation (such as removal of ectoparasites or because it "feels good"). The following equations describe Model II: F I F + B O - o D N+l N N N ON+1 = K1 DN+1 = K2’ if FN < “ O(FN-u), if FN > u, K1 > K2 Suppose the parameters stay the same as in Model I. a = 2 a = B = 1 u = 5 and K1 = 5 K2 = 2 The time courses of the values of FN’ ON’ and DN can be computed for each of the three experiments as was done for Model I. The initial values are as follows: FN = 10, ON = 4, DN = 2. Time FN ON DN 0 10 4 2 1 12 4 10 2 6 4 l4 3 0 4 2 4 2 4 2 5 4 4 2 J .. I. "e " ”E p.91. ”‘2‘. LJ.IJLEJ1-‘.'SE ."_ 211 t. i" A‘- s. Test Test Thus, with Model 11, a deprivation effect is Obtained for D 71 one day depriva- 8 4 tion 8 12 4 O 6 4 Five 8 4 days 12 4 of 16 4 dust 20 4 deprivation 24 4 6 28 4 N 0 38 as found in experiment 1 and there is an increase in FN as found in experiment 2. baseline Test What about experiment 3? Time FN ON 0 6 4 remove Oil gland one week of 8 0 recovery (depri- 8 0 vation) Access to dust 8 8 0 Access to dust 9 2 O Deprivation 10 O 0 Test 11 0 0 These values continue for all subsequent post-Operative tests. 0 2 ’12:: a W4 f. ."t-w.’ I77" ,I . ., The time course for values of D fairly closely to the actual results of experiment 3. that the value of DN declines to (or stays at) the level of K Alternative Starting Values 72 N as predicted by Model 11 correspond The Model predicts 2. To indicate the generality of Model 11, different starting values of FN - 10, Test Test Thus, with results of baseline Test N Time 0 one day of deprivation 2 0 Five days of dust deprivation 6 FN 10 8 12 10 8 12 16 20 24 28 O I 4 and DN I 6 will be used. 4 O 38 different starting values, Model II still predicts the both experiments 1 and 2. For experiment 3: Time FN 10 remove Oil gland one level of recovery (depri- vation) 8 8 73 8 0 0 Access to dust 8 8 0 6 Access to dust 9 2 O 6 Deprivation 10 O O 2 Test 11 O 0 2 These values continue for all subsequent post-operative tests. With these starting values, Model II again corresponds closely to the results of experiment 3, since the value of DN does not decline to zero, but to the value of K The degree of decrease in D from the base- 2' N line test 2 to post-Operative test 1 will depend, however, on the initial level of DN' A comparison between the two models is shown in Tables 7 and 8. Table 7 shows which experiments can be predicted by Model I and II. Table 8 traces a time course of values of F ON’ and DN when the two N, models are allowed to continue for a period of time. As can be seen, the values predicted by each model exhibit cycles, with more extreme fluctuations in values exhibited by Model I than Model II. lgplications for Further Research Table 7 shows that Model 11 predicts a less variable time course for values of DN than does Model 1. However, there is still change over time in the value of DN’ which suggests that the time span between reliability tests may be an important determinant of the reliability coefficient obtained for the frequency of dustbathing components. Model II also employs a constant (K2) which sets a lower limit to the value of D Experimental manipulation of K2 (perhaps through modifying the N. 74 Table 7 A comparison of predictability of Models I and II. Predicts results of Model I Model II Experiment 1 NO Yes Experiment 2 No Yes Experiment 3 NO Yes 75 Table 8 A time course of values of FN’ ON’ and UN for Models I and 11. Model I Model II Time FN ON DN FN ON DN 0 10 4 2 10 4 2 l 12 0 10 8 4 10 2 2 0 14 2 4 6 3 0 6 0 O 4 2 4 6 10 0 2 4 2 5 16 O 2 4 4 2 6 l4 0 22 6 4 2 7 0 0 18 8 4 2 8 0 10 O 10 4 6 9 10 10 0 8 4 10 10 20 0 10 2 4 6 ll 10 O 30 0 4 2 12 0 O 10 2 4 2 13 O 10 0 4 4 2 14 10 10 0 6 4 2 15 20 O 10 8 4 2 76 experience of the bird) would yield different time courses for D N after removal of the uropygial gland. Moreover, since K2 requires some minimum amount of dustbathing regardless Of the level of FN’ a systematic decrease in DN with removal of the uropygial gland would have been found in experiment 3 if very high levels of FN were present at the start of the experiment. Thus, perhaps presenting a deprivation period of 5 days (rather than 1 day) prior to the post-Operative tests would have led to more positive results. Also, the model is sensitive to slight changes in procedure. For instance, if only one day of access to dust were presented prior to the post-operative tests, the predicted results for experiment 3 would have been different. A 9...) ‘ Mun-‘9‘...“llwl tutu-Mm l l . 4 r ‘ l V H" ”'TITI'ITH’JLI’HMIMIflfiflflfgfijlflfljfljfuflfllim