Science & Technology

Begging for Milk: Evolution of Teat Massaging in Suckling Pigs

We present a model that simulates the evolution of teat massaging (begging) by suckling piglets (Sus scrofa) and milk distribution among teats (provisioning) by their mothers. Contrary to previous begging models, this one incorporates an ontogenetic
of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  J. theor. Biol.  (2002)  215,  321–332doi:10.1006/jtbi.2001.2511, available online at on Begging for Milk: Evolution of Teat Massaging in Suckling Pigs I. Dosta ¤ lkova ¤* w , M.  $ S pinka z  and P. Kindlmann* *Faculty of Biolo g ical Sciences ,  Uni  v ersity of South Bohemia ,  Czech Academy of Sciences ,  Brani  $ so v sk ! a 31,  CZ   37005  $ Cesk ! e Bud  $ ejo v ice ,  Czech Republic and   z Etholo g  y Group ,  Research Institute of Animal Production ,  CZ   140 00  Pra g ue - Uh $ r !  ın $ e v es ,  Czech Republic ( Received on  25  July  2001,  Accepted in revised form on  5  December  2001) We present a model that simulates the evolution of teat massaging (begging) by sucklingpiglets ( Sus scrofa ) and milk distribution among teats (provisioning) by their mothers.Contrary to previous begging models, this one incorporates an ontogenetic dimension in thatthe inherited begging and provisioning rules are repeatedly allowed to interact, and theirconsequences for milk intake, growth and death probability are assessed, during eachnursing. We test the model under three selection regimes differing in the relative importanceof the between-litter selection component. We show that the selection regime with thestrongest between-litter selection component leads to lowest begging levels and most effectivemilk utilization, thus supporting the hypothesis that selection based on whole litters mayattenuate sibling competition. r 2002 Elsevier Science Limited. All rights reserved. Introduction Food solicitation (‘‘begging’’) by young animals,which induces changes in parental provisioningbehaviour is common among altricial birds (e. g.Redondo & Castro, 1992; Price & Ydenberg,1995; Lotem, 1998; Leonard & Horn, 2001).Evolutionary interactions between the ‘‘beg-ging’’ function (level of solicitation),  x ð c Þ ;  of ayoung in condition,  c ;  and the ‘‘provisioning’’function,  y ( x ), describing the amount of foodsupplied by a parent in response to begging level, x ;  result in a ‘‘parent–offspring conflict’’: off-spring demanding more resources than it isoptimal for the parent to provide (Trivers,1974; Godfray, 1995). In the clutches of morethan one young, there is also ‘‘sibling competi-tion’’ with each offspring striving to get a largershare than its nestmates (Mock & Parker, 1998).In many mammalian species, a large part of teatmanipulation during nursing is non-nutritive(Hall & Williams, 1983; Lidfors  et al  ., 1994; $ Spinka & Algers, 1995; de Passill ! e & Rushen,1997). This manipulation affects milk release(Wakerley  et al. , 1988; Bruckmeier, 2001) andpossibly also milk production (Jensen  et al  .,1988; Svennersten, 1992; Jung, 2001), whichenables a tactile form of begging to evolve.Yet, this form of food solicitation has been muchless investigated than the visual and acousticsignalling of young birds to their parents.If the fitness of an offspring  f  ð c ; x ;  y Þ ;  dependson condition  c  of the young, its begging level  x and food received from the parent,  y ;  while theresidual fitness of the parent  g ð  y Þ  depends onlyon  y ;  then there exists an evolutionary stablestrategy, i.e. a pair of functions  x  ð c Þ  and  y  ð x Þ such that any modification of either of them will w Author to whom correspondence should be addressed.E-mail:$35.00/0  r 2002 Elsevier Science Limited. All rights reserved.  be selected against. At this ‘‘signalling equili-brium’’, young in lower condition beg more andparents provide more food in response to higherbegging (Godfray, 1991). Thus, begging carriesan ‘‘honest’’ information about the need of anindividual young and parents use it to allocatecare more efficiently. Yet, because of the siblingcompetition, this honesty is only achieved at acost: the begging itself consumes a substantialpart of the provided food. If the parentalresponse  y ð x Þ  is fixed and the begging  x ð  y Þ depends on the current provisioning rather thanon the long-term condition, then the ‘‘scramblecompetition’’ model applies (Parker & Macnair,1979; Mock & Parker, 1998). The model suggests that even higher levels of competitive beggingwill evolve especially in large broods, which maysubstantially reduce parental fitness.The above-mentioned analytical models F inorder to be tractable F do not incorporate manyfactors typical for realistic parent–offspringinteractions in natural populations, like feedingfrequency, typical growth curves and mortalityrates, mutual interactions between siblings, foodconversion efficiency and learning abilities.Numerical models have shown that these mightinfluence the begging/provisioning interaction.For example, Rodr ! ıguez-Giron ! es  et al  ., (1998)have shown that Godfray’s (Godfray, 1991)‘‘signalling’’ equilibrium is unstable under rea-listic conditions and populations slide into asecond, ‘‘non-signalling’’ Nash equilibrium withzero begging level. One important aspect thathas been ignored up till now is that interactionsof   x ð c Þ  and  y ð x Þ  occur more than once withineach generation, as young are usually fed severalhundred times before fledging or weaning(Redondo & Castro, 1992). To our knowledge,neither a model, nor a simulation of the offspringbegging/parental provisioning co-evolution as-suming its dynamics during ontogenesis has everbeen published.Here, we fill this gap by developing a modelthat simulates the evolution of teat massaging(begging) by suckling litters of piglets ( Susscrofa ) and milk distribution among teats(provisioning) by their mothers.  Sus scrofa  waschosen as a model species, because understand-ing the evolution of ‘‘begging-through-teat-massaging’’ has a potential practical application.In the herds of domestic pigs, the selection of new breeding females among the young animalsoccurs according to different criteria, includinglitter size at weaning, litter weight at weaningand individual growth performance. The breed-ing programmes thus put artificial selectionpressures on the amount of milk that the mothergives to her progeny and on the distribution of the milk. The pattern of milk provisioning, inturn, may significantly depend on teat massaging(see below), and understanding the co-evolutionof teat massaging and milk provisioning is thusimportant for our ability to predict the outcomesof breeding programmes.We test our model under three selectionregimes, which differ in the relative strength of the ‘‘individual selection’’ vs. ‘‘group selection’’components. That is, in one of the regimes theselection of animals, which will be bred in thenext generation, was based on individual pigletswhereas in the other two it was based partly orfully on the condition of whole litters. It hasbeen suggested for animals that live in stablegroups and compete for a limited resource thatpositive selection for individual performancetrait (such as growth) can have a negative effecton the mean performance in the population(Muir, 1997). This is because the increase in theperformance of the individual has, throughcompetitive behaviour, a negative influence onthe performance of its group mates. Viewing thisfrom the opposite perspective, selection at thegroup level should decrease competitive beha-viour. Thus, one can assume that in stable-group-living animals with restricted resources,group-based selection will result in larger and/orfaster increase in the mean performance thanselection based on individuals (Wilson, 1997).Empirical evidence supporting this hypothesishas been found for  Tribolium castaneum  (Wade,1997) and domestic laying hens (Craig & Muir,1996; Muir, 1997). For litters of mammalianyoung, this hypothesis specifically predicts thatbetween-litter selection for performance traitslike weaning litter size or weaning litter weightmay act against sibling competition, attenuatebegging and other competitive behaviour andimprove the utilization of the resource, i.e.the milk. Assuming that begging-through-teat-massaging is a competitive behaviour, we test I. D OSTA Ł LKOVA Ł  ET AL. 322  this hypothesis here: we test, whether regimeswith stronger between-litter selection componentwill lead to lower begging intensity and moreeffective milk utilization. We assess our predic-tions by addressing the following questions:(i) How will evolution shape the begging andmilking functions and will they differ betweenthe selection regimes?(ii) How efficiently will energy contained inmilk be used for piglet growth under the threeregimes? Which proportion of milk will beconverted into biomass of piglets that surviveduntil weaning as opposed to that spent onbegging and/or invested in piglets that diedbefore weaning?(iii) How will mortality and averagepiglet weaning weight develop under the threeregimes? Biological Background—Suckling in the Pig Female pigs give birth to litters of 3–8 (wildboar, Mohr, 1960) or 6 –14 (European domesticbreeds) piglets that each weigh over 1kg at birth.Within the first 1–3 days of life, each piglet formsan attachment to one teat of its mother andsucks almost exclusively from that teat untilweaning (Rossillon-Warnier & Paquay, 1984; dePassill ! e  et al  ., 1988; Puppe & Tuchscherer, 1999).Piglets are nursed in about 1hr intervals duringdistinct nursing periods lasting several minutes,but the actual milk intake occurs only during 15– 20s of milk ejection period in the middle of nursing (Fraser, 1980). During the rest of thenursing period each piglet noses, rubs and (non-nutritively) sucks its own teat. The intensity andduration of this teat massage (‘‘begging beha-viour’’) is enhanced if the piglets are hungry orare gaining weight too slowly ( $ Spinka & Algers,1995). There is some evidence that a moreintensive massage increases the amount of milkreceived in future nursings ( $ Spinka & Algers,1995; Jensen  et al  ., 1998), although it is unclearas to whether the effect is central or local.Massaging consumes extra energy since pigletslose weight at a faster rate during the non-nutritive phase of the nursing than when resting(Klaver  et al  ., 1981; Noblet & Etienne, 1986).Piglets suffer large pre-weaning mortality(10–25% in domestic herds) during the first daysof life (Daza  et al  ., 1999; Herpin  et al  ., 2001).The low-weight piglets are at a much higher risk(Weary  et al  ., 1998; Marchant  et al  ., 2000; Roehe& Kalm, 2000) and this early loss of low-prospect progeny may be a case of adaptivelitter size reduction (Forbes, 1993; Fraser  et al  .,1995). In European domestic herds, weaning ismost frequently imposed between Day 21 andDay 35 post-partum. Different criteria areapplied for selecting new breeding animals,depending on the main purpose of the herd,e.g. whether replacement animals are raisedwithin the herd or purchased from outside,whether the herd sells breeding animals or justanimals for slaughter, whether the main purposeof the selection is genetic improvement infertility or in growth rates. The Model We simulated the evolution of the begging andprovisioning functions. Begging function,  B , wasassumed to be the level of udder massage eachpiglet performs, dependent on its weight, relativeto its littermates. Provisioning or ‘‘milking’’function,  M  , was assumed to be the amount of milk that the mother releases to individualpiglets in response to their relative levels of ‘‘begging’’. The basic model structure is shownin Fig. 1 and its parameters in Table 1. We simulated a series of   k  hypotheticalbreeding generations. Each generation consistedof   N   animals (sows), each having  n  offspring(only female offspring were considered). Littersize at birth, nursing frequency, the total amountof milk given per nursing and mortality wereassumed to be the same in all animals andgenerations. We arbitrarily assumed 11 relativeweight categories of the piglets at each instant.Each individual,  I  , (both sow and piglet) wascharacterized by its birth weight,  W  (0), and two11-dimensional vectors,  B  and  M :  I  = I  ( W  (0),  B , M ) ¼ ( W  (0), ( B (1),  B (2), y ,  B (11)), ( M  (1),  M  (2), y ,  M  (11))). Here  B ( i  ) is the intensity of begging,provided that the individual belongs to the  i  -thweight category and  M  (i) is the relative amountof milk that this individual provides as aresponse to begging of an offspring belonging– from this individual’s point of view–tocategory  i  . These vectors represented the begging EVOLUTION OF TEAT MASSAGING IN SUCKLING PIGS  323  and provisioning functions, which were ‘‘inher-ited’’ and ‘‘evolving’’ in a way described below.The weight of each suckling piglet,  W  ( t ), wasassumed to be a function of time,  t .The structure of the simulations is shownin Fig. 1. They consisted of a set of nestedcomputer procedures. The main one, called E  v olution , consisted of repetitive usage of procedures  Reproduction ,  Life  and  Selection .Procedure  Life  simulated the processesfrom birth to weaning in each sow and herlitter F growth of piglets, their begging andmilking by the sow. Procedure  Selection  simu-lated the process of selection of a new generationof sows from the offspring of the previousgeneration of sows. Finally, procedure  Repro-duction  simulated the process, in which a pigletselected for further breeding will give birth to heroffspring F thus it described the weights andinherited characteristics (begging and provision-ing functions) of each of her offspring. Theinitial values used in the model are defined inTable 1. We now describe each of theseprocedures. Procedure Reproduction : We arbitrarily as-sumed that each mother gives birth to  n  piglets, I  i  ,  i   ¼ 1, 2, y ,  n , each with its own character-istics,  I  i  ¼ I  i  ( W  i  (0),  B i  ,  M i  ). The birth size,  W  i  (0),was chosen from a normal distribution (seeTable 1 for parameters). Components of vectors B i   and  M i   were defined as B i  ð i  Þ ¼  B ð i  Þ þ  R b B v ar ð 1   j I     r b j = 10 Þ ; M  i  ð i  Þ ¼  M  ð i  Þ þ  R m M  v ar ð 1   j I     r m j = 10 Þ ; where  B v ar  and  M  v ar  are parameters,  r b  and  r m were integers chosen from a uniform distributionin the interval  o 0;11 4  and  R b  and  R m  werechosen from the normal  N  (0,1) distribution. ReproductionBeggingAllocationGrowthDeathSelection:Litter weightIndividual survivalLitter size − individual weight Fig.  1. Flow diagram describing the basic structure of the model. Table  1 Parameters used in the model  Parameter Notation Type Constant value Initial valueNumber of mothers  N   Constant 10  F Litter size at parturition  N   Constant 8  F Piglet birth weight mean  W  0 ( 0 ) Constant 1.5kg  F Piglet birth weight variation  F  Constant 3 variance  F Begging  B  Evolving  F  0.002 for all 11 weight categoriesMilking  M  Evolving  F  0.004* k ,  k =0,1, ..., 10  k  is categoryHeritability of begging 1/ B v ar  Constant 60  F Heritability of milking 1/ M  v ar  Constant 90  F Number of nursings until weaning  K   Constant 600  F Milk per nursing  k T   Constant 0.2kg  F Milk to body weight conversion  C   Constant 0.2Probability of death  P  Dependenton weight F F Number of generations Iter Constant 100  F I. D OSTA Ł LKOVA Ł  ET AL. 324  Thus the values of   B i  ð i  Þ  and  M  i  ð i  Þ  were similarto those of the parent F they were ‘‘inherited’’,and the begging and provisioning functionsdiffered from those of the parents in that at arandomly chosen category (defined by  r b  and r m ), they were ‘‘disturbed’’ and this ‘‘distur-bance’’ faded out with the distance from thiscategory. The sizes of the disturbance wererandom, but their means were proportional to B v ar  and  M  v ar  –thus the inverses of these para-meters might be considered as a measure of ‘‘heritability’’ of the begging and provisioningfunctions. Procedure Be gg in g : At any instant and withineach litter, piglets were categorized into 11categories according to their weight relative totheir sibs. The category,  c ð t ; i  Þ ;  of   i  -th piglet attime  t  was defined as c ð t ; i  Þ ¼  5  þ  int  20  W  i  ð t Þ P n j  ¼ 1  W  i  ð t Þ = n    1 !"# with additional conditions: if   c ð t ; i  Þ  4 10, then c ð t ; i  Þ ¼ 10 and if   c ð t ; i  Þ o 0, then  c ð t ; i  Þ ¼ 0, where W  i  ð t Þ  is the weight of   i  -th piglet in the litter of total size  n  piglets (see Table 2). The energy spentby begging of this piglet, was then set equal to B ð c ð t ; i  ÞÞ : Thus, piglets with less then 75% of averagebody weight were put into category 0, pigletswith 125% and larger weight into category 10,piglets with average weight into category 5, etc.(Table 2). Thus, this procedure determines thebegging intensity that the piglet exerts as afunction of its own weight, relative to its sibs,and to its inherited ‘‘strategy’’ for begging F thevalue  B ð i  Þ :  While begging and provisioningfunctions are ‘‘inherited’’ and fixed during juvenile life, the category of the piglet isdetermined at each time step and therefore maychange according to how it grows relative to itssibs. Procedure Allocation : Similar to weight cate-gories, at any instant and within each litter,begging of piglets was also categorized into 11categories. The category,  b ð t ; i  Þ ;  of begging of   i  -thpiglet at time  t  was defined as b ð t ; i  Þ ¼  5  þ  int  20  B i  ð t Þ P n j  ¼ l   B  j  ð t Þ = n    1 !"# with additional conditions: if   b ð t ; i  Þ 4 10 ;  then b ð t ; i  Þ ¼ 10 and if   b ð t ; i  Þ o 0, then  b ð t ; i  Þ ¼  0 :  Asthe begging functions  B ð i  Þ  were not necessarilydirectly proportional to  i  ;  begging of a piglet didnot necessarily correspond to its weight cate-gory. Each piglet was then allocated the amountof energy  M  al  ð t ; i  Þ  contained in milk, which madeup the proportion  M  ð b ð t ; i  ÞÞ  of the total amountof milk,  k T  ;  offered by the sow to her litterduring one nursing: M  al  ð t ; i  Þ ¼  k T  M  ð b ð t ; i  ÞÞ P n j  ¼ l   M  ð b ð t ;  j  ÞÞ Procedure Growth : The growth of piglets frombirth to weaning was modelled as W  ð t  þ  1 Þ ¼  W  ð t Þ þ  c ð M  al  ð t Þ   B ð t ÞÞ ; where the time unit was equal to one between-nursing interval and  c  is the energy to bodyweight conversion ratio. Thus, the increment inweight per unit time was assumed to be propor-tional to the difference between the energygained from milk and energy spent by begging. Procedure Death : At each instant, each pigletwas assumed to die with the probability P  ¼  max ð 0 ; 1    A ð W  i  ð t Þ   0 : 5 Þ B Þ ; where  A  and  B  are constants, in the model weused  A  ¼  0 : 97 ; B  ¼  0 : 07 : Table  2 Range of the weight categories Category  W  i  ð t Þ P n j  ¼ l   W   j  ð t Þ = n 0 (–  N ;  0.75 S 1 (0.75,0.8 S 2 (0.8,0.85 S 3 (0.85,0.9 S 4 (0.9,0.95 S 5 (0.95,1.05)6  / 1.05,1.1)7  / 1.1,1.15)8  / 1.15,1.2)9  / 1.2,1.25)10  / 1.25,+ N ) EVOLUTION OF TEAT MASSAGING IN SUCKLING PIGS  325
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!