Monday, October 28 -- Altruism

[Topic List] | [Next Lecture] | [Previous Lecture] | [Discussion]

Last time we spoke of kinship and dispersal, which lead to altruism. To avoid inbreeding, one sex usually disperses at maturity. Usually it's the males who disperse since benefits of dispersal are more to them than to females, since they get greater mating opportunities. Also, the costs are less; access to food is reduced which isn't as important to them as to females. When males don't leave, then females are forced to. This happens especially in the new world primates. So, because of different patterns of dispersal, you have variations in which sex is living with relatives. Whoever lives with relatives has the opportunity to treat relatives differently than non-relatives. Called kin-correlated behaviors, these include spatial proximity, agonistic helping, and grooming. So they treat relatives differently, but why? Can we explain it in terms of natural selection? This is for today.

So how do Primates know kin from non-kin?

You can't treat your relatives differently if you don't know who they are.
There are two hypotheses:

Social learning

You grow up with your parents so you know who they are since they're always right there. Imprinting! Lots of evidence and research on this. Probably extended farther in primates since they have more extended relationships. You learn who your mother is, then pay attention to who she hangs out with and that teaches you who your other relatives are. Some problems: You can make mistakes or get fooled. Like cuckoos- trick other parents into taking care of their egg. Question: how do cuckoos avoid trying to mate with the kind of bird that raised them if they imprint on their mom?? Another problem is how to know who your dad is in non-monogamous societies. So people hypothesized another theory- this is not based o observations, just on people sitting around thinking.

Self-referential phenotype matching

What does this mean? All your physical characteristics as opposed to genotype which is all the genes which went into guiding your development. So if you see someone who looks like you then maybe they're related and maybe you should treat them better. This isn't just visual, but also smell or sound or something. Dawkins' armpit effect- imagining people sniffing another's armpit and then their own and deciding that they're related or not. Out text says this does exists but more recent publications are really getting to doubt it. One study where it was demonstrated on macaques and one in ground squirrels but hey have both been criticized as to methods and no on has been able to repeat their findings.

Natural Selection and the Expectation of Selfish Behavior

Remember the characteristics of natural selection:
1 phenotype variation that ...
2 affects reproductive success and...
3 is due at east partly to genetic differences

You expect individuals to act selfishly. It is easy to understand and explain through natural selection. This is the easiest way to ensure that you will have the highest reproductive success. By simple mathematics, genes for altruism will die out.

There are four ways of behavior to another- it might benefit or harm you, an it might benefit or harm the other guy.


Benefits recipientHarms recipient
Benefits donorcooperationselfish behavior
Harms donoraltruismspite

Selfish behavior, we expect to see. Cooperation we can understand. Spite is hard to explain in terms of natural selection, but that's ok because you almost never see it. However, in animals you see altruism but how can we explain it?

An example- grooming. It benefits the recipient, but what does it do for the donor? Another example is helping defend a group against a predator; sometimes male colobus will confront predator instead of running away with the rest of the group. In primates there are many examples of individuals taking care of infants which aren't their own. Another example is alarm calls.

A Counter-example of Apparently Altruistic Behavior: Alarm Calls

If the predator moves slowly, alarm calls don't cost much- just a few calories. But with quick predators, if you call out the alarm then that might catch attention of the predator who might come after you.

Group selection interpretation
This was the classic explanation; it was good for the group. As people sharpened up their logic, they debunked this theory. George Williams especially drove home the point that if you have a bunch of altruists it's really easy for a cheater to arise and take over.

Question; how to resolve the paradox?
How can altruistic behavior evolve in individuals within the framework of natural selection? The key insight in cracking this problem was realizing that it's not individuals who are passed down but genes. If your relatives are like you then they carry the same genes.

Two solutions were developed to answer this question: kin selection and reciprocal altruism, which is less well established.

Remember the basics of natural selection: If there's a trait in a population which spreads, then it's favored by natural selection. Now we will complicate things by adding inclusive fitness. We are going beyond thinking about individuals and are now thinking about the genes that they carry. Alternative forms of genes, called alleles, are in us. We can copy our genes not only by having offspring ourselves, but by helping our relatives to have offspring. You can calculate exactly what is the probability that another individual will have the same genes as you.

First, what is the probability that a copy of a particular gene in a parent is shared by one of its offspring?
(Thanks to William Hamilton)

Remember:


The probability of sharing any particular gene can be computed as follows:
At each generational link between two individuals, there is a meiotic event, resulting in a 50% probability that a copy of a particular gene is passed on. For T generational links, this probability is:

(0.50)^T


(Note that our text uses L instead of T)

To compute r, you just sum this value for all possible pathways between 2 individuals. It can get a little more complex because you can be related to the same person in two different ways. Here are some examples;

Parent and offspring:
One link, parent to child. Thus, r=.5

Grandparent and grandchild:
Two links; child to parent, and parent to grandparent. (.5)^2 = r =.25

Half siblings:
Two links; up to the parent is one and back down to the other sibling is another. (.5)^2=.25, so your relatedness coefficient is .25

Full siblings:
These are a little tricky because there are two ways that their genes could be the same. Full siblings are related through both their mother and through their father, so each pathway is .25 and when you add them together, it's .5.

Two cousins:
Their parents are siblings, so there are two pathways with four links each that they're connected by. 2 * (.5)^4=.125

Genetically speaking, full siblings are just as related to you as your offspring. So an individual can benefit just as well by helping its siblings as by making offspring.

If given the choice to help a grand-offspring or a half-sibling (both have r=.25), which will an individual be better off choosing? They will usually choose the younger one because the benefits t the recipient will be greater since a young helpless infant will be more helped by a single act of helping than an old established animal. (Also, lifetime reproductive potential is greater for younger individuals than for older ones.)

Kin Selection

Hamilton noted that selection would favor behaviors or traits that not only increase one's own reproduction, but also that of others, insofar as those others are related and share genes with you. This form of selection was termed kin selection to contrast it with Darwinian natural selection.

To formalize the concept, Hamilton introduced the concept of inclusive fitness, a measure of one's own reproduction plus a summed component of the reproduction of others, devalued by the degree of relationship.

Inclusive fitness
=
individual fitness
+
fitness of others devalued by
the degree of relationship between donor and recipient

Conditions favoring altruism

Whether or not you will help someone depends on thee factors:
How related you are
How much it benefits them
How much it costs you

Hamilton's rule is for predicting whether or not altruistic behavior will be exhibited:

B/C > 1/r
ORrB>C
ORrB - C > 0

B = benefit to recipient
C=cost to donor
r = coefficient of relatedness of donor to recipient

A Return to Alarm Calls

The theory would be that even though the alarm call has a cost, the benefit to kin must be more than the cost to the caller. So, if you can measure the costs and benefits, you should be able to predict whether or not an animals will call.There have been plenty of studies on this, especially done on ground squirrels. It turns out that the rate of alarm call depends on closeness of relatives and how many are nearby.

In ground squirrels, males disperse and don't live with relatives, but females stay so they're with more relatives. Based on their percentage of the population, females call much more than you'd expect and males much less.

Male gorillas are related to all the kids in the group, but females are only related to a few of the kids. So you'd expect males of do more alarm calls than females and they do.

Also, vervets were studied in captivity. They checked whether females did more alarm calls when offspring were present, and they did indeed call much more when offspring were present than when there were others present.

A cautionary note: To understand all this, you have to do a lot of math. You may be thinking, "Do monkeys really understand all these concepts and everything?" But you must understand that the animals don't have to be conscious beings for it to work, and they don't have to know calculus! Natural selection does all the math for them- if an individual's behavior works within the formulas, then its genes will be passed on, but if its behavior isn't within the formulas, then the overly- or underly-altruistic genes won't be around for long.

Note that there are different kinds of altruism:
Phenotypical altruism, which has a cost to the individual but is also genetically selfish because it benefits the genes that the individual carries. This is explainable by natural selection.
Genetic altruism, on the other hand, harms the individual and doesn't perpetuate any of the individual's genes, and so it can't be explained by natural selection.


This kin selection theory explains most acts of altruism that we see- they're generally benefitting relatives. This is in contrast to:

Reciprocal Altruism

(Developed by Robert Trivers in the 70's) This concept will be easy for us as humans to understand since we do it a lot. The idea is that the altruism is reciprocated at some other time- an individual does something nice and later gets it back. This has gotten a lot of theoretical discussion but there is little evidence.

Primate examples (From pg 287 in coursepack)
Our first example is baboons: Remember their consortships- when a female is receptive, a male will follow her around and keep others away as long as he can. Sometimes the individual in the consortship will be the highest male in the group and will be able of keep everyone away. But sometimes two males will gang up to drive him away and then one of them will mate with her. They often recruit an ally to help chase away the other guy. The helper incurs serious costs, but what benefits does he get? He doesn't get a female. You notice through further observations that in the future the helper will recruit the guy he helped to help him drive away a guy and get a female.

Our second example is vervets: Here, alliances are between females.When a fight between females breaks out, one participant will often scream for help and other individuals who hear her will come to her aid. Usually, it's relatives who come but sometimes it's non relatives. How likely is it that a non-relative comes to give help depends on how recently the recipient groomed the donor. Data from playback experiments shows that when kin hear a scream for help, they look around for the source for a long time when there was no prior grooming, while non-kin look for the source for a LOT longer when they had been recently groomed, and barely looked at all when they hadn't been recently groomed.

Conditions favoring reciprocal altruism
To illustrate this point, we played prisoner's dilemma. When you're playing a series with the same partner, the 'tit-for-tat' strategy works the best. When playing with a different partner every time, it's better to defect every time.

To learn the theory of prisoner's dilmma, check out this link:
http://newciv.org/GIB/BI/BI-36 .HTML

So, how favored by natural selection a tit-for-tat strategy is depends on the likelihood of future interactions with the same individuals.

Conditions necessary for the evolution of reciprocal altruism




Wednesday, October 30 -- Cooperative Polyandry

[Topic List] | [Next Lecture] | [Previous Lecture] | [Discussion]
Cooperative Polyandry: a Definition
"Groups in which two or more males mate with a single female during a single breeding season and collaborate to raise her offspring"


Expectation of Polygyny and Selfish Behavior

Cooperative polyandry seems to fly in the face of expectations in two ways- First of all, males are expected to raise their reproductive success by gaining access to females. Females don't maximize reproductive success by getting more males! Secondly, it looks like altruism- one of the males is taking care of offspring not related to him! Why is he being a sucker?? Cooperative polyandry is pretty rare for these reasons.

It occurs in scattered animals- some raptors like hawks and eagles have polyandrous mating systems. Also african hunting dogs. In primates, humans in Asia with traditional cultures have polyandrous mating systems, usually for much the same reasons as callitrichids do, who are the only non-human polyandrous primates.

Only a few have been studied; common marmoset, cottontop tamarin, golden tamarin, saddleback tamarin.

Cooperative Polyandry in Callitrichids

Characteristics of callitrichids
They're the smallest of anthropoid primates; all under 1kg. Generally territorial. Diet is mixed- fruit and small prey items. Some eat gums, especially the marmosets who are specialized gummivores. Another important feature is that they produce twins- about 80% of births are twins. They're only primates who twin routinely.

Results of early studies
When people first started looking at social systems and mating systems, they thought callitrichids were monogamous. This was based on a few things- there's only one breeding female per group in the wild and also polyandry was pretty unknown at that point. Furthermore, in captivity the monogamous groups seemed to breed more successfully and when you added adults of either sex you ended up with a lot of aggression. However, when people got out into the wild, looked at things in more detail, and paid better attention, they figured out how things really work.

Saddleback tamarins: mating and communal care systems
Saddleback tamarins group composition: of all groups observed, how many were which composition type?

22%One male, one female
61%Multiple males, one female
14%Multiple males and multiple females
3%All males

Just because their group composition is mostly polyandrous doesn't necessarily mean that they're mating polyandrously. In many animals, there are males and females in a group together but there's only one breeding male- so the social system is not the same as the mating system. For instance, in red-tailed monkeys there is more than one male in the group but only one of them is mating with the females.

Are tamarins really polyandrous? Let us look at data from a group with two males in it. They each got about half of the mating (60-40% split for male1 and male2). Not only were the matings shared, but so was the parental care. Looking at statistics for infant carrying, male1 carried infants 45% of total carried time, male2 about 37%, and the mom carried less than 20%. We know that one of the males is being altruistic because only one of them could have been the father. He is being a non reproductive helper. These are seen a lot in birds- there are often helpers at the nest, usually young from previous seasons, such as in bee-eaters. Also jackals do this. Some primates have non-reproductive helpers too, like the barbary macaques.

Helpers and Helping

Do helpers actually help?
In some cases, the results have been mixed. In some studies of birds, the nests with helpers really do better, but in others they don't seem to. Some behaviors look to us like they would be helpful but you have to look closely to see if they actually increase reproductive success. In tamarins, having help does actually help increase reproductive success. This help is mainly gotten from polyandrous males. You also get helping from other non-reproductive helpers in the group- usually older siblings. Some data on carrying from a monogamous callitrichid group:

Family MemberPercent of Carrying Time
Mom23%
Dad35%
Female sibling #125%
Female sibling #213%

Why do helpers help? What do they get out of it?

"The key to explaining helping and polyandry in tamarins is the high need for parental care.
Only one of 33 sets of offspring in the wild was born to a single pair. This one exception was conceived in a trio of 2 males and one female.
Thus, monogamous pairs do not attempt to breed, and helping appears to be necessary for successful reproduction."

Ok so we know that the NEED is high, but does this really provide the motivation for the helpers? No, of course not! We need to know how it benefits the helper. Remember our equation rB>C. We know that helping is absolutely necessary. This means that the benefits to the recipient are really high- thus, if there is even a little bit of relatedness, then it will be worth it to help.

Costs of parental care
High neonate:mother weight ratio

The average weight of newborns in all primates is 8% of the mother's body weight, but in callitrichids it's 18% of the mother's body weight.

Costs of lactation
When they're pregnant, females spend 25% of their time feeding, but when they're lactating, they spend 35% of their time feeding. Since the babies are growing so quickly, they're expending a lot of energy and mom is still providing it all.

Costs of carrying
Look at how time budget is affected when moms are carrying infants around- they need to spend more time resting when they're carrying children around.

ActivityPercent of time budget when carrying kidsPercent of time budget when not carrying kids
resting9258
feeding330
travelling312

So you can see that carrying children around really affects how much they can get done.

The Evolution of Cooperative Polyandry

"Helping is essential for successful reproduction, and the callitrichid mating system is determined by the number of nonreproductive helpers.
1 monogamy occurs in groups where a sufficient number of older offspring remain to help.
2 polyandry occurs in groups lacking old offspring. In these situations, pairs accept/recruit the help of an additional male."

Basically, males who are mated benefit under some conditions by accepting another male into the group and allowing him sexual access to the female. If the first male is trying to raise kids alone with the wife then it's just too hard. So he can have sole access but no surviving kids, or else he can share with another male and get half the kids sprung. Then if there get to be older siblings, the first male doesn't have to accept the secondary male any more.

So, monogamous pairs are the ones who have several older offspring in the group. Monogamous pairs who don't have older offspring don't attempt to mate. This is the evidence that backs up the previous theory and explains why males would tolerate another male in their group.

What's in it for the older offspring? They seem to be making the best of a bad situation. The habitat is often saturated, and when an offspring reaches the age when it could disperse, it can't because all the available spaces are occupied and defended by older, more experienced animals. They probably could leave the group, but they may as well stay since they're safer from predators. So they do the best they can of the situation; if they can't have their own kids they may as well help raise siblings who have some of their genes. They may also benefit by learning how to parent, gaining experience on parenting skills.

Monogamous and Extended Family Groups of Marmosets

There is some more recent evidence noting the differences between tamarins and marmosets. It seems that tamarins are more polyandrous and the marmosets are more into monogamy with non-reproductive helpers.

Why? It seems that the need for polyandry is less in marmosets. Their home range size is smaller on average than tamarins', as well as their day rage length. This means that they don't have to travel so much, so carrying infants around isn't as much of a chore.



Friday, November 1 -- Allomothering

[Topic List] | [Next Lecture] | [Previous Lecture] | [Discussion]

Again, we're looking at particular topics in primatology because they have a special interest. This is interesting because the current theory on natural selection works on individuals, not groups, and we don't see too many behaviors that are good for someone else. This is one of these behaviors which looks like it's good for someone else, so it's very interesting to behavioral ecologists.

What is it?

"Non-maternal infant care." In other words, infants being taken care of by anyone who isn't their mother.

Who are potential allomothers?

One real obvious case is in the monogamous species; the dad. Another case is non reproductive helpers like in the callitrichids. Another is females other than the mom. In some cases, this female alloparent isn't even closely related to the mom and baby. In any species, infants are re real focus of interest. Sometimes there are certain signals just for infants. For instance, in barbary macaque, there's a special grin used basically only in greeting infants.

Care by monogamous males and pre-reproductive individuals

In monogamous species, it's pretty easy to see how the males gain by helping since they have a pretty high confidence of paternity and their opportunities for other matings are limited.

We can understand why non-reproductive males in polyandrous systems would help, as we talked about last time.

We can also understand why older siblings would help through the theory of inclusive fitness, because they are increasing their own reproductive success by helping out their siblings.

However...

Care by non-maternal females

This is not rare- it's seen in a number of old world and new world species, particularly in species where the females don't disperse. Vervets, cebus monkeys, squirrel monkeys, and macaques are all known for allomothering.

One thing they help out with is carrying. This is probably the biggest thing they do to help. They also sometimes provide food for offspring who aren't necessarily their own. Alloparents also guard them from predators or potentially hostile conspecifics. One form which is quite rare is nursing infants who aren't their own; even when a female is looking after an infant who isn't her own, she generally won't let it nurse. However, in cebus monkeys it has been observed on occasion.

The earliest observations of alloparenting behavior were back in the 40's in rhesus monkeys, in captivity. They documented allomothering by adult females who would carry around babies but weren't always nice to them.

Jane Lancaster showed that young nulliparous females were doing disproportionally much of the allomothering in vervets. In colobines like the leaf monkeys, sometimes an infant will be taken by another female right after it's born- when it has barely dried- and it will be passed from one female to another through the group, and the mom seems totally fine with it.

Moms obviously get a benefit but what are the costs? Babies don't always get treated great by the alloparent- sometimes there's a tug of war between two females to see who gets to hold the baby. Sometimes the alloparent doesn't carry it correctly- you see babies dragged around by their foot, sat upon, dropped. Also, sometimes the allomother will abandon the baby. These are risks to the baby and through it to the mom. So two questions arise;


  1. Why do allomothers do it? Why should allomothers devote time and energy to care for infants of other individuals?
  2. Why do mothers permit it? Why should mothers tolerate non-mothers handling their infants given the risk of abuse?

We will tackle each question in term, but first some terminology;


parity= state of having had offspring before
nulliparous=not having borne offspring
primiparous=having borne one offspring
multiparous=having borne more than one offspring

What benefits do caretakers accrue?

The learning-to-mother hypothesis
This is the best-supported hypothesis. Parenting skills are learned and allomothering helps to learn these skills. As Jane Lancaster noticed, primates start breeding late in life, have a long gestation period, and in general only produce one infant at a time. If they lose an infant through inexperience, it's a real setback. It's a real reproductive benefit to learn how to be a better mom.

Predictions Predicted by this Hypothesis

  1. Caretakers should be primarily nulliparous females- since they're the ones who need the experience.
  2. Allomothering should promote mothering skills.

Evidence
In langurs, it is true that nulliparous do show more interest in young infants than experienced females do. Sara Hrdy, who studied langurs for many years, wondered whether females were indeed more interested in allomothering before they'd have their own offspring. She found that nulliparous females tried to allomother more frequently than what you'd expected based on their proportion of the group's population, while parous females tried it much less than expected from their population in the group.

There is also data from vervets; based on their percentage of population, juveniles do disproportionally much of the allomothering compared to older females.

The other question is, does it actually help the allomothers become better mothers? Lynn Fairbanks studied vervets and found that first-time mothers with high alloparenting experience raised 100% of their first offspring to maturity, but mothers with low experience had less than a 50% survival rate of their first infant.

Is kin selection involved?
You may be saying, 'Can we really dismiss kin selection so easily? It's so common; are we sure that's not what's responsible for the behavior of allomothers?' Basically, yes, they're sure. The main reason they don't think it's involved ;

"Kin selection is not implicated as a factor in the origin of allomaternal care.
Juvenile female vervets initiate caretaking about at same rates whether (3.5 bouts per hour) or not (3.6 bouts per hour) they had infant siblings in the group.
Female langurs manifest an interest in infants from other groups."

In fact, sometimes when females from two groups get together, females from one group kidnap a baby from the other group and hold it for sometimes hours despite the efforts of the mother to get it back.

Of course, kin selection is around; when given a choice, the number of caretaking bouts with siblings were much higher than expected given the number of nonsibs available compared to siblings available. This leads us to conclude that the kin directed part of this behavior is secondary to other motivations. Also, older kids don't always prefer their siblings.

What do mothers gain?

Especially in the first few weeks of life, infants get a lot of attention from everyone in the group. In most cases, the mom lets the infant go. So what's in it for her?

Hypothesis:
Mothers benefit through a reduction of time and effort allocated to infant care.

Predictions:

Are the predictions supported by the data? Yes.
There is data from vervets, langurs, and squirrel monkeys that suggests that females feed faster and more efficiently when allomothering is available to them. All either feed longer or more efficiently when someone else is holding their infant.

The second prediction is also supported: When you compare the amount of time that a female's infant is carried by allomothers and how much time there is between her births, you see that as the amount of allomothering support goes up, the interbirth interval goes down. Thus, mothers who have babysitters can eat more, so they save up resources more quickly and can give birth again sooner.

Another idea which doesn't have as much data support: If another female has been allomothering for you and you're killed, then it's more likely that your infant will be adopted. Sort of like cultivating potential godmothers. Most evidence simply suggests that it frees mom up to forage more and reproduce faster.



Discussion --

[Topic List] | [Next Lecture]

Today we turned in papers, got back papers, and reviewed the lectures.


Let me know your thoughts: phyl@umich.edu
Last modified: November, 1996