Tietelbaum data

[From Bill Powers (950804.1145 MDT)]

Bruce Abbott (950804.1120 EST) --

Tietelbaum's qualitative assessment doesn't agree with your quantitative
analysis. Tietelbaum:

     When few bar presses are required to obtain each pellet, as with 1
     and 4 bar presses per pellet, all animals press the bar and obtain
     quantities of food which essentially reflect their ad libitum food

Your numbers show a steady decline of food intake with increases in
ratio. The only way to create two different regimes is to draw an
arbitrary line and say that above the line the food intake is
"essentially the same as the ad libitum intake" and that below it

     ... all animals fail to press the bar enough to obtain sufficient
     food to maintain a stable weight level. (p. 488)

In fact, the rate of food intake simply decreases as the ratio
requirement goes up, in a smooth and significantly continuous way. When
you fit your straight line and deduce the collection time and behavior
rate, you get the same result as before: the behavior rate is constant,
not a function of the ratio or the food intake.

     Unfortunately Teitelbaum does not report any of the weights. One
     thing is clear from the description of procedure: the rats were not
     exposed to the schedules long enough to produce significant weight
     loss. They remained at each ratio value only two days. At the
     lowest ratios they apparently were able to maintain their weights;
     at 64 and 256 they lost weight, but this would have occurred only
     over the last fours days of the study. At these ratios
     (particularly at 256), perhaps the energy expendature required to
     earn a pellet was hardly worth the return. It may also be that
     these lab-reared rats had sufficient stores of excess body fat that
     by the end of the experiment they were only down to levels
     approaching those of rats in the wild. In other words, there was
     not yet much error in the nutrient-control system because the
     system was buffered by the animal's store of fat.

You should know by now what I think of ad-hoc explanations like these.
Teitelbaum _did not report the weights._ So you know nothing about
whether the rats maintained their weights on the low ratios, and thus
you don't know whether reliance on the store of body fat occurred at the
high ratios, etc., etc.. Talking about "significant" weight loss implies
that weights were measured and the losses were deemed insignificant. We
don't know what the losses were. Your explanation concerning the use of
stored body fat implies that there WAS a weight-loss; you don't burn
body fat without losing weight. But we don't know: the necessary data
weren't reported.

     Perhaps the kind of regulation you envision does not come into play
     until the rats are operating without the reserve of excess body fat
     common in cage-reared laboratory rats. Until that time, other,
     short-term factors come into play such as stomach loading, pleasure
     derived from the taste of the food, and so on. Meal size and
     frequency may be products of other control systems, whose
     references are only slowly altered by long-term changes in error in
     the nutrient control system, which accumulate on a time-order of
     days rather than minutes or hours.

I am baffled by these numbers.

     I think they make sense in the context of the data and analysis I
     provide above.

The data you provided about the rats in question (not the VMH rats or
the quinine rats or rats in different studies) do not answer my
questions at all. The only _numerical_ analysis you provided showed a
steady decline of intake with increased ratio, and no apparent attempt
by a rat to increase its behavior to prevent the decline, during a two-
day period in which (at the extreme ratio) the food intake had fallen by
7/8. Stomach loading, pleasure from taste, and so on must have also
declined in frequency or amount as the intake declined.

I am presuming, from your remarks yesterday and the present description,
that the pellets were delivered one at a time and were 90 mg in size.
According to the _Joy of Cooking_ (one of my technical reference
manuals), protein and carbohydrate average about 4 large calories per
gram. Teitelbaum, in a slightly different reference I have in my old
notes, reports that rats consume from 50 to 60 large calories per day
when free-feeding, which comes out to 12.5 - 15 grams per day. At FR-1,
the animals were receiving 132 * 0.09 = 11.8 grams per day, which is on
the low side of free-feeding intake at least for the average rat (if
Tietelbaum had reported the free-feeding consumption for each rat, we
wouln't have to try to deduce it indirectly). At FR-64, they were
receiving less than 1.5 grams per day, which can hardly be explained as
the natural diet of a wild rat. A normal rat maintains its body weight
by consuming only slightly less food than a VMH rat maintaining three
times the body weight. Permanently reducing food intake to 1/8 of free-
feeding intake would surely kill a rat.

For this reason I haven't bothered to ask whether the wasted food was
weighed. One possible explanation might be that the rats really needed
only 1.5 grams per day, and simply ate that amount off each pellet
before dropping the rest through the floor of the cage. But I think we
can discount that, unless we find a similar effect during bouts of
eating (Tietelbaum reported that rats typically consume their daily
intake in about ten distinct meals).

I think that we can draw only one conclusion from these data: the rats
in these experiments showed no short-term control over ingestion that
relied on varying the rate of behavior. Because body weight was not
reported, we don't know if they had any long-term control over body
weight, either. By your theory, the actual rate of behavior was not
influenced by any variable in the experiment: it was constant.


I assume that it's not self-evident to a rat that if it can get food by
moving an object up and down repetitively, it can get more food by
moving it up and down at a faster rate. This relationship has to be
learned before it can be used for controlling food intake. Clearly,
these rats had not learned that relationship. If they had been trained
on a variety of schedules, and had been given enough experience to
convert an intake error into a variation in pressing rate, we might have
seen a very different outcome.

This means of getting food does not naturally occur; rats have to learn
it, if they can learn it at all. This may be too complicated for a rat.
All the evidence we have so far shows that it is. Perhaps rats have
reached their limits in learning either to press a bar repetitively or
not to press it, with deprivation affecting mainly the speed with which
they maintain repetitive pressing over a long period.

As it is, in these and other experiments you've analyzed, the only
possible control system left is one that exerts very slow long-term
control over nutritional level or body weight by varying the general
level of activity. And even to verify that, we would have to do
experiments that lasted long enough for a control process to reach a
steady state.

So far, it seems, we have not found any evidence of control in any of
these experiments.

Bill P.