The Imminence of Behavioural Change and Startle Responses of Chicks

THE IMMINENCE AND STARTLE OF BEHAVIOURAL RESPONSES OF CHANGE CHICKS by A. D. CULSHAW and D. M. BROOM') 2) (Department of Zoology, University of Reading, Reading, England) (With 9 Figures) (Acc. 5-VIII-1979) 1. INTRODUCTION The motivational state of an individual is likely to change during a bout of activity, the degree of change varying according to the activity. If of nutrient level there that activity serves a function such as maintenance is a maximum likely duration of its continued occurrence before repletion some of the causal factors necessary is reached. As repletion approaches, for the occurrence of eating must be declining and the probability that an This would not be the reason for eating bout will end will be increasing. to others, that is to say the termination of an activity which is subordinate the of gaps is determined one whose occurrence entirely by appearance to at least is dominant activities. Eating, however, between dominant some other activities, for example in doves (MCFARLAND, 1974). Even in the case of a subordinate upon causal activity, or an activity dependent factors which do not change during the course of a bout, motivational state is still changing during a bout because other causal factors will be is interrupted, the probability Hence, if a bout of behaviour changing. behaviour will replace the ongoing behaviour that any particular may change as the bout progresses. The previous paper (FORRESTER & BROOM, 1980) describes the startle are chaged by responses of 6-day-old chicks when their surroundings of a torch-bulb on the wall of the home pen. The the illumination or has been ocresponses differ according to which activity is occurring, The chicks to that the 5 seconds prior change. compared curring, during in the experiments reported here were all engaged in one of two activities for helpful discussion. A. D. C. was supported by a 1) We thank R. C. FORRESTER Science Research Council studentship. 2) Reprint requests should be sent to D. M. BROOM. 65 but, at the time of testing, different amounts of time had elapsed since the of 6-day-old chicks which had beginning of the bout. The responsiveness with that of chicks for which the terstarted was feeding compared just mination of the feeding bout was imminent. Similar experiments were carried out on chicks which were preening at 2, 4, and 6 days of age. When describing a collection of samples of behaviour it is most useful to express the durations of the behaviours, or of the intervals between them, logarithmically (NELSON, 1964, SCHLEIDT, 1965). This has been done in this study in the determination of the minimum inter-bout interof val by means a log survivor curve (NELSON, 1964; WIEPKEMA, 1968; DELIUS, 1969; van RHIJN, 1977). The method is described in detail by SLATER (1974, 1975). This is not the only way to assess bout-length (MACHLIS, 1977) but was used as it provides a convenient means of classifying points in bouts into those where the bout is likely to continue, e.g., of 2-3 seconds after the start, and those where there is a high probability imminent transition to another activity. II. METHODS The methods were similar to those described in the previous paper (FORRESTER & BROOM,l. c. )with the following exceptions. The eggs were incubated throughout development in our laboratory. The chicks were observed for 5 minutes, before startling by torchbulb illumination and for a further 10 minutes whilst the torch-bulb remained illuminated. The 3W torch-bulb was fixed to the rear wall of the pen when startling chicks which were feeding. This position of the bulb was then similar, in relation to the chick, to that at the front of the pen when the chicks were preening. Other experiments indicated, however, that the position of bulb had no effect on responsiveness (CLJLSHAW, 1977). The statistical tests are 2-tailed Mann-Whitney U tests. Experiment1. In order to determine bout-length, 15 undisturbed chicks were watched for several hours on days 2, 4 and 6 and their behaviour recorded. Experimenl2. Chicks aged 6 days were watched by a silent observer through a one-way screen and their behaviour was recorded until feeding occurred. The observer continued to record behaviour in detail but moved a switch after feeding had continued for 2 seconds. For 15 chicks this resulted in immediate illumination of the bulb but for 15 others the bulb was illuminated after a delay, determined from the bout-length measurements, so as to conincide with the predicted end of the bout. For these latter chicks, recording was terminated if feeding stopped before bulb illumination. The decision as to whether to illuminate the bulb at the beginning or the end of a bout was always taken prior to any recording. Each chick was tested once only, either at the beginning or at the end of a bout. None of the chicks had seen any illumination change since being placed in the pen from the darkened incubator. Experimenl3. The methods were the same as in Experiment 2 except that the chicks were preening the wing. 15 chicks were startled at the beginning of the bout and 15 at the predicted end of the bout on day 2, day 4 or day 6. The lengths of preening bouts were calculated separately for the three ages. 66 ' 1. Determination III. of RESULTS bout-lengths. was to compare the effects of interSince the aim of these experiments chick behaviour at an or a late stage in a bout of feeding, or rupting early of preening, it was necessary to determine the mean bout-length for these activities. The sequences of pecks during feeding are separated by brief we wished to distinguish such gaps gaps. In order to measure bout-length from the intervals between bouts of feeding. There are sometimes short from long gaps intervals between bouts which are indistinguishable within bouts. When analysing undisturbed behaviour it was necessary to establish some criterion to be used in order to decide that a bout had ended. This was done by plotting log survivor functions, of gap or interval length for feeding and for preening. In these log survivor functions, of gaps of a particular the log of the frequency length is plotted in a the If the lengths of cumulative fashion negatively against gap-length. then the function would have constant gaps were distributed randomly to the probability of occurence of any slope, for the slope is proportional If the distribution of events is the log clustered, particular gap-length. survivor curve will show a steep initial drop. If the probability of a further event, after a cluster of events, is low and does not change with time then the function will have a shallow unchanging slope after the initial steep drop. The discontinuity point between the steep drop and the shallow, unchanging slope gives a useful indication of the length of the shortest interval between bouts. Since it was not possible to obtain enough data from each chick, during one day of age, to determine its feeding and preening bout-length, data chicks and used in Experiments were collected for 15 undisturbed 2 and were measured for 3. All gaps between continuous feeding movements each of the chicks at 6 days of age. In total, 377 gaps were measured and the log survivor function is plotted in Fig. 1. The first and sharpest is at 10 seconds, after which there is a constant, lower prodiscontinuity bability over at least four points. A sequence of feeding not followed within 10 seconds by further feeding was called a bout and the lengths of all bouts on the same records of behaviour were measured. For 49 bouts of feeding the mean bout-length was 83.8 seconds. Our criterion for a point at which the ending of the bout, and change to another activity, was imminent was therefore established as 84 seconds. Bout termination is much more likely at this stage in the bout than after 2 seconds. 67 Fig. 1. For feeding, the logarithm of the total number of gaps remaining after successive subtraction, is plotted against the length of gap in seconds. The line is drawn to emphasise the position of the shallow unchanging slope after the initial steep drop. The discontinuity point at the beginning of this slope is used as the critical gap-length in determining when a bout has ended. (see text for further explanation.) were collected for Similar data on gaps between preening movements and mean length of 15 chicks on days 2, 3, 4, 5, and 6. The distribution gaps did not vary with age (CULSHAW & BROOM, in prep; CULSHAW, 1977) so all 1198 gaps were used when compiling Fig. 2. The steep initial slope in the log survivor function persists until 6 seconds, after which there is a at 22 seconds. constant slope until another, much smaller, discontinuity a critical of 6 the mean seconds, lengths of bouts were Using gap-length 11.9 seconds at 6 days (n = 125), 10.2 sec at 4 days (n = 89) and 8.4 seconds at 2 days (n = 108). 68 at beginning or end feeding bouts (6-days-old). when a chick is startled Feeding is an activity which is inhibited (BROOM, 1969). As expected, therefore, it occurred rarely during the first two minutes after the illumination in the chick's home of the torch-bulb The of at this time was less than 5 $lo of that frequency pen. feeding pecks which undisturbed chicks would have shown during a comparable period after starting a feeding bout (CULSHAW & BROOM, in prep). Chicks startled at the end of a feeding bout (84 seconds after the start) did not resume Those feeding during the 10 minutes that the bulb was illuminated. startled at the beginning of a bout (2 seconds after the start), however, were more likely to feed at a low level during minutes 3-10 after bulb illumination (Fig. 3). 2. Interruption Fig. 3. The mean number of pecks at food per min. by 15 6-day-old chicks, is plotted for successive 20 sec. periods during 5 min. whilst undisturbed and during the 10 min. after the illumination of the bulb on the wall of the pen. Key: · stimulus on (bulb illumination) at the beginning of bout of feeding, o-stimulus on at end of bout of feeding. The most obvious aspects of the startle response of young chicks after are a period of reduced activity, associated with fixbulb illumination the followed bulb, ating by a period of increased activity which may be associated with loud peep calls and jumping as well as fixating (BROOM, the first type of 1966, 1969; FORRESTER & BROOM, 1. c. ). In this experiment all birds but was shown few showed much increase in response by locomotor activity or loud calling. As shown in Figs 4 and 5, the proportion of time fixating the bulb or immobile during the first 5 minutes after bulb illumination was clearly greater amongst chicks startled at the end of the bout of feeding than amongst those startled at the beginning. Undisturbed chicks which have started to feed are rarely recorded fixating the bulb or immobile but occasionally crouch and often walk or peck at non-food objects during the next 5 minutes. (FORRESTER & BROOM, I. c. ; 69 CULSHAW & BROOM, in prep). In Experiment 2, chicks startled at the of a and closed their eyes bout sometimes crouched beginning feeding minutes 5-10 after bulb illumination. during They were slightly more to less to walk or than chicks startled at the end peck likely peep and likely of a bout. None of these differences was large. The area within 5 cm. of the bulb was avoided by those startled at the end of the bout (0-60 seconds after illumination 0-5 minutes and 5-10 minutes p<0.01, Hence these chicks also avoided the p < 0.05). adjacent food container. 3. Interruption at the beginning or end of preening bouts (2, 4, or 6-days-old). Preening the wing, like feeding, is an activity which is inhibited when a chick is startled (BROOM, 1969). Undisturbed chicks are likely to preen within & 5 minutes again (FORRESTER BROOM, 1. C.;CULSHAW & BROOM, in Fig. 4. The mean number of seconds that the 15 6-day-old chicks fixated the bulb during each successive 20 second period is plotted for 5 min. whilst undisturbed and for 10 min. after the illumination of the bulb. Key (as Fig. 3): *-beginning, o-end. First 5 min. after bulb illumination beginning v. end p < 0.05. Fig. 5. The mean number of seconds that 15 6-day-old chicks were immobile during each successive 20 sec. period is plotted for 5 min. whilst undisturbed and for 10 min. after the illumination of the bulb. Key (as Fig. 3): ·-beginning, o-end. First 5 min. after bulb illumination beginning v. end - p<0.05. 70 occurred amongst prep) but at each of the three ages, little preening a chicks startled at the end of preening bout (Fig. 6). Whereas chicks startled at the end of a bout ceased preening almost immediately, those to preen for startled at the beginning of a preening bout often continued some seconds so that their preening frequency was considerably higher during the first minute after bulb illumination. Fig. 6. The mean number of seconds that the 15 2, 4, and 6-day-old chicks preened their wings during each successive 20 sec. period is plotted for 5 min. whilst undisturbed and for 10 min. after the illumination of the bulb on the wall of the pen. Key: ·-stimulus (bulb illumination) at beginning of bout of preening the wing, o-stimulus on at end of bout of preening. First min. after bulb-illumination, beginning v. end, 2 days p < 0.01, 4 days p<0.001, 6 days p<0.01. Chicks which are preening are very seldom recorded during the next 10 minutes fixating the bulb, immobile, or peeping. They do not walk feed or crouch (CULSHAW & much in that time but they do sometimes BROOM, in prep). Chicks startled at the end of a bout of preening fixated the bulb for longer at each age (Fig. 7) and were immobile for longer at 4 and 6 days (Fig. 8). There were no large differences, related to the timing of bout interruption, for other measures of behaviour. Walking and jumminutes after were more 5-10 bulb illumination, ping frequent, slightly amongst 2 and 4-day chicks startled at the end of the bout (Fig. 9). Some of these chicks showed stereotyped locomotor movement from side to side at the front of the cage. Activities which were more likely amongst chicks 71 Fig. 7. The mean number of seconds that the 15 2, 4, and 6-day-old chicks fixated the bulb during each successive 20 sec. period is plotted for 5 min. whilst undisturbed and for 10 min. after the illumination of the bulb. Key (as Fig. 6): *-beginning, o-end. First min. after bulb illumination, beginning v. end, 2 days p<0.01, 4 days p<0.01, 6 days p<0.01. Fig. 8. The mean number of seconds that the 15 2, 4 and 6-day-old chicks were immobile during each successive 20 sec. period is plotted for 5 min. whilst undisturbed and for 10 min. after the illumination of the bulb. Key (as Fig. 6): *-beginning, o-end. First min. after bulb illumination, beginning v. end, 4 days p < 0.01. 72 Fig. 9. The mean number of seconds that the 15 2, 4 and 6-day-old chicks walked during each successive 20 sec. period is plotted for 5 min. whilst undisturbed and for 10 min. after the illumination of the bulb. Key (as Fig. 6): *-beginning, O-end. of the bout included startled at the beginning seconds p < 0.05), preening and eyes shut. IV. crouching (day 2, first 60 DISCUSSION The general result of these experiments is that chicks startled at the end of a bout of feeding or preening the wing show a greater startle response behaviour than are chicks and are less likely to continue with undisturbed startled at the beginning of a bout of feeding or preening. Two activities which are characteristic of the response and rare in disturbed chicks, fixwith the eyes open, are clearly more frethe bulb and immobile ating quent amongst chicks startled at the end of a bout. The other component of the chick's startle response involves increased activity and sometimes conditions and shows no peeping. It is less frequent in these experimental This clear difference according to the point of behaviour interruption. of the fact is further evidence for the idea that there are two components and escape (BROOM, 1969). response, immobility The differences between the startle responses shown by chicks which are feeding, preening etc., are described by FORRESTER & BROOM (loc. cit. ). In our experiments feeding and preening were suppressed during the 10 minutes after bulb illumination but feeding did occur if chicks were of a feeding bout and preening did continue startled at the beginning briefly if chicks were startled at the beginning of a preening bout. Chicks startled at the end of a bout were immobile and fixated the bulb at first, then started walking and pecking at non-food objects if they had been pecking previously at food. If they had previously been preening they did 73 not peck but walked and showed stereotyped movements similar to the of movements adult fowl which have been "thwarted" pacing (DUNCAN & differences in response which WooD-GusH, 1972). These are qualitative vary according to ongoing behaviour and the timing of the interruption during the bout. The differences in fixating the bulb and immobility are in time-course. but there are some differences quantitative principally startled the end of a of are for much of at bout immobile Chicks preening but for little time after that. This the first minute after bulb illumination, was not shown All chicks startled at the chicks. pattern by 2-day-old rather than at of a of the bout end, beginning, preening showed less imoften became immobile time but some after the bulb illuminamobility tion. In these experiments, as in those of FORRESTER & BROOM (l. c. ), the behaviour sequence of the chicks is interrupted by a change in the chick's which evokes a clear but not an extreme surroundings response. A drastic to environmental would be initiate extreme changes in change likely motivational state so that the response would depend little upon the previous state. These results do provide, however, clear evidence that the motivational state changes during the course of an activity. at beginning and end of a bout The comparisons between interruption were similar for chicks of each age. The younger chicks showed less marked responses, as has been reported previously (BROOM, 1969), and there were some differences in the pattern of their response but the chicks startled at the end of the bout showed a greater as well as a different response at 2, 4 and 6 days. In addition to the difference in immobility mentioned of the response was the above, a difference in the time-course in 6-daymore prolonged suppression of walking after bulb illumination old chicks than in those aged 2 or 4 days. in responsiveness are commonly explained in terms of acVariations that atmotivation and Our results demonstrate attention. tivation, to all in behaviour on the basis of an activation tempts explain changes model such as that of DUFFY (1962) are simplistic. Feeding is usually in activation continuum placed high any (MORuzzI, 1969) so it is likely that the activation level would fall during a feeding bout if it changes at and sleeping (CULSHAW & is often followed by crouching all. Preening BROOM, in prep.) and these are generally considered to be low activation so activation should often decline during a preening bout. behaviours, should be Responses greatest when activation is high so simple activation models would predict lower responsiveness at the end of a bout of feeding The reverse was found in these exor preening than at the beginning. 74 Hence, within the time scale and activity range of this study, periments. there is little support for an activation model. In explaining our results we consider that one of the ways in which A motivational variables. systems operate is by modifying attentional useful way of describing motivational state is as a point in causal factor space (SIBLY & McFARLAND, 1974). At the beginning of a bout of activity, the causal factors necessary for the occurrence of that activity are at a initial part of an aclevel. increase the They may during relatively high mechanism. Adult of a feedback terns or gulls because positive tivity which have started to preen are more likely to stop preening and change to another activity in the very early part of a preening bout than in a bout which has lasted for a third of the mean bout-length (VAN IERSEL & BOL, levels of some relevant 1958; VAN RHIJN, 1977). As the bout progresses, of the activity. causal factors are likely to decline, partly as a consequence At the same time, causal factors necessary for the occurrence of other activities may be increasing. The bout of activity would normally end, either when causal factors necessary for a second behaviour increased to a or inlevel which resulted in that behaviour being shown (competition time or when second hibition the behaviour was during sharing) McFARLAND disinhibited & SIBLY, 1975). (MCFARLAND, 1974; We postulate that the change in levels of causal factors during the latter stages of a bout initiates an attentional change. Since the animal is carrya feeding or preening bout it ing out the same type of activity throughout is likely that some attentional variables remain the same throughout the of the sensory receptors and the databout. For example, the orientation relevant to this activity may be similar at different processing procedures variables as the end of a stages of the bout. The changes in attentional bout approaches include an increased may ability to attend to sensory inwhich is not related to the put directly ongoing activity. It is likely that in the of a bout of behaviour falls first few seconds as the distractibility on" "lock mechanism (FORRESTER & BROOM, I.c.) comes into effect. The reverse change occurs towards the end of the bout. Such a change in could explain our finding that responsiveness is greater distractibility nearer to the end of a bout than 2 seconds after the beginning. DELIUS the of inthat rate and level information (1970) suggested processing such a creases as a point of transition mechanism Perhaps approaches. facilitates the changeover from one activity to another and increases the that it will occur at the optimum time. Animals whose probability behaviour is interrupted near the end of a bout are more prepared for than they response to any sort of change in input and less constrained, 75 were at the beginning of the bout, by the effect of high levels of the causal factors necessary for the ongoing behaviour. Further experiments using these techniques should provide insight into the control of greater behavioural sequences. SUMMARY The responses of chicks startled near the end of a bout of feeding or preening were different from, and measurably greater than, those of chicks startled at the beginning of bouts of these activities. This result provides evidence about changes in motivational state during a bout of activity. It is proposed that causal factors, which are changing during the bout, may initiate an attentional change as the point of transition to another activity approaches. Distractibility and information-processing rate may be increased near the end of an activity because the individual can then attend to a greater variety of inputs than it could earlier in the bout. This would maximise the efficiency of the ensuing behavioural change. In order to carry out these experiments, bouts of feeding and preening were measured using a criterion based on log survivor curves for inter-event gap-lengths. The effects on the response of the time during a bout at which interruption occurs, were shown by chicks of 2, 4 and 6 days of age. The method used in these experiments, the interruption of activities at previously defined instants whilst observing behaviour in detail, affords opportunities for studying the nature of behavioural sequences and their control. REFERENCES BROOM, D. M. (1966). Changes with age in the domestic chick's reactions to novelty. - Anim. Behav. 14, p. 586-587. -(1969). Reactions of chicks to visual changes during the first ten days after hatching. - Anim. Behav. 17, p. 307-315. CULSHAW,A. D. (1977). The development of responsiveness in domestic chicks. Ph. D. thesis, University of Reading. DELIUS,J. D. (1969). A stochastic analysis of the maintenance behaviour of skylarks. Behaviour 33, p. 137-178. -(1970). Irrelevant behaviour, information processing and arousal homeostasis. Psychol. Forsch. 33, p. 165-188. DUFFY,E. (1962). Activation and behaviour. - Wiley, New York. DUNCAN,I. J. H. & WOOD-GUSH,D. G. M. (1972). Thwarting of feeding behaviour in the domestic fowl. - Anim. Behav. 20, p. 444-451. R. C. & BROOM,D. M. (1980). Ongoing behaviour and startle responses FORRESTER, of chicks. - Behaviour 73, p. 51-63. IERSEL,J. J. A. van & BOL, A. C. A. (1958). Preening of two tern species. A study of displacement activities. - Behaviour 13, p. 1-88. MACHLIS,L. (1977). An analysis of the temporal patterning of pecking in chicks. Behaviour 63, p. 1-70. Adv. MCFARLAND,D. J. (1974). Time sharing as a behavioural phenomenon. Study Behav. 5, p. 201-225. -- & SIBLY,R. M. (1975). The behavioural final common path. - Phil. Trans. R. Soc. Ser. B. 270, p. 265-293. MORUZZI,G. (1969). Sleep and instinctive behaviour. - Archs ital. Biol. 197, p. 175-216. NELSON,K. (1964). The temporal patterning of courtship behaviour in the glandulocaudine fishes (Ostariophysi, Characidae). - Behaviour 24, p. 90-145. 76 RHIJN, J. G. van (1977). The patterning of preening and other comfort behaviour in a herring gull. - Behaviour 63, p. 71-109. SCHLEIDT,W. M. (1965). Gaussian interval distributions in spontaneously occurring innate behaviour. - Nature Lond. 206, p. 1061-1062. D. J. (1974). A state-space approach to motivation. - In: SIBLY,R. M. &MCFARLAND, Motivational control systems analysis. ed. MCFARLAND, D. J.), Academic Press, London. SLATER,P. J. B. (1974). Bouts and gaps in the behaviour of zebra finches, with special reference to preening. - Rev. Comp. Animal 8, p. 47-61. -(1975). Temporal patterning and the causation of bird behaviour. - In: Neural and Endocrine Aspects of Behaviour in Birds. (WRIGHT,P., CARYL,P. G. & VOWLES,D. M. eds). Elsevier, Amsterdam, p. 11-33. WIEPKEMA,P. R. (1971). Positive feedbacks at work during feeding. - Behaviour 39, 266-273. ZUSAMMENFASSUNG Die Reaktionen von Küken, die gegen Ende eines Fress- oder Putz-'bouts' aufgestört worden waren, waren anders und messbar grösser als diejenigen von Küken, die zu Beginn der Anfälle dieser Aktivitäten gestört worden waren. Dieses Ergebnis gibt einen Hinweis auf Änderungen des motivationellen Stadiums während eines Aktivitäts-'bout'. Es wird vorgeschlagen, dass kausale Faktoren, die sich während des Ablaufens der Aktivität ändern, einen Umschwung in der Aufmerksamkeit bewirken, wenn sich der Zeitpunkt des Übertritts zu einer anderen Aktivität nähert. Die Ablenkbarkeit und die Geschwindigkeit der Informationsverarbeitung kann am Ende einer Aktivität erhöht sein, da das Individuum dann auf eine grössere Vielfalt von Inputs achtgeben kann als es zu einem früheren Zeitpunkt konnte. Dies würde die Wirksamkeit der darauffolgenden Verhaltensänderung maximieren. Um diese Experimente durchführen zu können, wurden Fress- und Putz-'bouts' jede für sich gemessen, wobei Kriterien für die Längen dieser Aktivitäten verwendet wurden, die sich auf 'log survivor'-Kurven der Intervalle zwischen dem Auftreten dieser Verhaltensweisen beziehen. Die unterschiedlichen Reaktionen, gemäss des Zeitpunktes der Unterbrechung, wurden von 2, 4 und 6 Tage alten Küken gezeigt. Die Methode, die in diesen Experimenten angewandt wurde, nämlich das Unterbrechen von Aktivitäten zu vorher definierten Zeitpunkten, während das Verhalten im Detail beobachtet wurde, bietet Möglichkeiten an, die Natur von Verhaltenssequenzen und deren Kontrolle zu studieren.