RE: "PACK HUNTING" THEROPODS

[Nathan Myhrvold <nathanm@MICROSOFT.com>]

Sorry for the late reply - was busy.  Comments inserted below.

> -----Original Message-----
> From: Larry Dunn [SMTP:larrydunn@hotmail.com]
> Sent: Monday, March 30, 1998 7:49 AM
> To:   dinosaur@usc.edu
> Subject:      Re: "PACK HUNTING" THEROPODS
> 
> From: Nathan Myhrvold <nathanm@MICROSOFT.com>
> 
> >Coordinated group foraging behavior does NOT require a lot of 
> >intellegence.
> 
> It does when discussing vertebrates.  Only the most intelligent 
> vertebrates do it.
[Nathan Myhrvold]  

There is no evidence for this proposition in several ways:

-  Many species of vertebrate predators are social living, ranging from
European badgers to the more famous examples of wolves and lions.   They
span nearly the full range of vertebrate predator sizes and intellegence
ranges.   Aquatic cooperative hunters range from fish to aquatic mammals.
Even within the mammals there is a very wide range.

-  There is no evidence that solitary vertebrate predators are any less
intellegent than social or group living/group hunting species.


> > If you doubt this, watch film of army ants bringing down
> >prey MUCH larger
> >than themselves, in what seems to be a coordinated fashion.   
> >Theropods
> >certainly had the potential to be as smart or smarter than ants!
> 
> Ants and other hive invertebrates aren't illustrative here.  It's 
> totally apples and oranges. Each organism in a hive is not a separate 
> identity but part of a constituent whole.  It's a beautiful arrangement 
> but totally alien to vertebrate behavior.  
[Nathan Myhrvold]  

I don't mean to be rude, but your answer  is incorrect on every basis:

-  There is nothing that prevents vertebrates from having a similar social
arrangement. Group living in a manner highly similar to the eusocial insects
has been reported for several vertebrate species, notably the naked mole
rat.   

-   It is true that eusocial insects are highly related to one another.
This explains the evolutionary MOTIVATION for extreme degree of cooperation
to evolve.   You have to separate two issues - the evolutionary motivation
that would lead to cooperative hunting, which is one issue, and the level of
intellegence (cognitive ability) required to get "cooperation" in hunting
and foraging.   Quite apart from the issue of WHY individuals would
cooperate, and why evolution would lead them to, you can ask what the level
of intellegence is for a particular behavior.  As it turns out, cooperative
foraging and hunting does not require much intellegence at all.   Ant are a
valid example in that specific sense.  Their overall social structure IS
different, but that is a separable issue.   

-  More generally, cooperative foraging does not seem to require more
intellegence than solitary foraging.  Leopards are smarter than ants, and
their style of hunting requires more intellegence.  However it is not clear
that lions or other cooperative hunters are smarter than leopards, nor it is
true that there is any obvious intellegence difference between solitary
foraging ants (of which there are many species) and those that hunt in
groups.

-  As to the evolutionary motivation, there is plenty of basis for this.
Group living / group hunting mammilian predators are typically genetically
related to almost the same degree as the eusocial insects - indeed they are
as closely related as mammilian reproduction allows.   For example, the
female lions in a pride are typically all sisters, or at most half sisters.
Prides typically have between one and three male lions who are typically
brothers.  Eusocial insects achieve even higher degrees of relation than
sisters because of differences in their reproductive genetics, however,
cooperative hunting mammilian predators come as close as they can.  With
these high degrees of relatedness, there is ample evolutionary motivation
for the evolution of cooperation.

To return the discussion to dinosaurs, a group of closely related theropods
could have the evolutionary motivation to develop social behavior and
cooperative hunting.  They have more than enough intellegence.


> Example: soldier ants throw themselves into the jaws of the warriors of 
> a competing hive to protect larvae.  Why?  They're expendable:  a 
> warrior ant does not represent a great investment of group resources.  
> Wolves will not similarly end their lives to protect cubs of the pack --  
> like all mammals, they may fight hard to protect their young, but 
> self-preservation keeps them from parental suicide.  
> 
> To emphasize why hive insects aren't illustrative here, try to imagine a 
> wolf throwing itself into the jaws of a grizzly bear to protect *another 
> pack member's* cubs.
[Nathan Myhrvold]  

Again, I don't mean to be rude, but there is no support here:

-  The main issue in "self sacrafice" is the question of reproductive
strategy.  Some creatures breed like mad and have high mortality.  Others
have smaller birth and death rates.   This is known as the r-strategists
versus K-strategists.   Insects - social or solitary - do die more often.
If you created fully eusocial vertebrates that were large bodied, long lived
creatures, then they would risk themselves, but not with the abandon that
ants do.

-  Wolves DO die, and risk death to defend pack members.  Cubs in a pack are
exclusively (or nearly so) the offspring of the alpha male and alpha female,
so for most of the wolves in the pack, they are *another* wolves cubs.
Pack members are often siblings, so they have some relation to the cubs,
which explains why they are willing to make these risks. The same is true of
other social mammilian carnivores - the can and do die protecting cubs.

There are many other examples.  African hunting dogs will reguritate food to
feed cubs that are not their own.  

Therepods that are sufficiently related could indeed have made "altrusitic"
sacrafices for one another.   So called "selfish genes" can make for
apparently altruistic phenotypes. This effect is independent of the degree
of intellegence of the creatures involved.

> >Many of the highly intellegent "pack hunting" mammilian carnivores 
> (wolves,
> >lions etc.) actually do far less actual coordination and planning than 
> once
> >thought.  
> 
> [BIG SNIP]
> 
> First, I agree that many have anthropomorphized pack hunting, but that 
> doesn't require that we then take the argument too far in the opposite 
> direction.  
> 
> If necessary, pack animals will in fact assist each other in dispatching 
> the prey animal.  If smaller prey is scarce, lions will attack a water 
> buffalo, and they pile on -- often six or seven lionesses are necessary 
> to bring one adult water buffalo down. 
> 
> Furthermore, certain degrees of herding strategies are involved.  Wolves 
> are the best example, but lions strategize to some degree as well.  And 
> watch hyaneas attempt to separate a rhinoceros calf 
> from its mother. 
> 
> In any event, pack hunters go out *together* to hunt down prey 
> animals -- singly or not -- to feed the entire pack.  To broaden the 
> definition of pack hunting as you've proposed because some people have 
> exaggerated the amount of herding and the frequency of assisted killing 
> would be to blur the very real differences in behavior between, say,  
> lions and oras.
[Nathan Myhrvold]  

It is far from clear whether the herding, or piling on behavior is really
all that "cooperative" in a deep sense.   In particular:

-  It does not necessarily require more intellegence than solitary hunting.

-  Individual action on the part of each so called "cooperator" can mimic a
wide range of "cooperative" behavior without any communication or
cooperation actually taking place.

The latter point is particularly important.  Schooling fish, most behaviors
of ants and many other animal behaviors are the COLLECTIVE result of
INDIVIDUAL algorithms.   What appears to be highly orchestrated - has no
score, and no composer.  Instead it is what happens when a bunch of
individuals following the same set of rules get together.

There are differences between the social behavior of lions and oras, but
that example does not make the point.  

> >Theropods could have been group living and group hunting without true
> >cooperation.    All they would need is social tolerance, and the 
> ability to
> >pursue individual forging strategies - like ants do. Arguments about 
> their
> >intellegence as a reason to doubt theropod pack hunting are thus rather
> >suspect.
> 
> Arguments about intelligence are critical in analysing group behavior 
> (as discussed, the hive-insect analogy is inapposite).
> 
> To reiterate, there is a direct correlation between intelligence and 
> cooperative hunting in vertebrates.  Pack hunting by its very nature 
> requires cooperation, not simply toleration.  Theropods may have 
> congregated on an animal brought down by one theropod (any real evidence 
> of this?) and they may have barely tolerated each other at kill sites 
> (they don't even seem to have done that, however), but that does't 
> constitute pack hunting.
[Nathan Myhrvold]  

Your comments reflect the common wisdom, but in fact are not supported by
detailed field study.

Cooperative hunting certainly spans a range of behaviors - from tolerance at
kill sites (as oras do to some degree) to deeper level of cooperation.  Even
among mammals there is a range of degrees of cooperation.  Most of the high
intellegence, deep cooperation stories are due to anthopomophic anecdotes.
The true degree of cooperation even in mammalian examples is far lower that
commonly supposed.

Evidence from extant animals shows it is entirely possible that closely
related groups of theropods COULD have a degree of what we casually term
"pack hunting".    It is difficult to say whether it was "as cooperative" as
a current mammals because behaviors do not fossilize!   However, it is NOT
impossible that the degree of cooperation was quite high - sufficiently that
we would recognize the behavior as "pack hunting" at some level.   

Nathan
>  
>