Introduction : The level of
sexual behavior displayed by rhesus macaque
males declines significantly with old age
(Phoenix, 1977; Phoenix and Chambers, 1982a;
Robinson et al., 1975), but serum testosterone
(T) levels, bound and free, do not decline
(Chambers et al., 1981). Nor do serum levels of
dihydrotestosterone (DHT) or luteinizing hormone
(LH) decline, and diurnal patterns of these
hormones do not differ in young and old rhesus
males (Chambers and Phoenix, 1981; Chambers et
al., 1982). if levels of sexual behavior decline
but T levels do not, T must be less effective in
activating sexual behavior in aging males.
Eventually, the level of sexual behavior
displayed by old males should not differ from
that shown by old, castrated males as T becomes
increasingly ineffective in supporting high
levels of sexual activity.
Fully adult (about 10-yr-old), sexually
vigorous rhesus males show a rapid depletion of
serum T levels after castration (Resko and
Phoenix, 1972) and a much slower but significant
decline in sexual behavior (Michael, 1972;
Wilson et al., 1972; Phoenix, 1973b). One can
restore sexual behavior to mean precastration
levels in at least 90% of such males by treating
them with testosterone propionate (TP; Phoenix
et al., 1973). However, TP does not increase the
levels of sexual behavior displayed by older
intact males (about 15-yr-old) whose sexual
behavior has declined (Phoenix, 1977), and TP
(10 mg/day) increases the sexual behavior
displayed by old castrated males (18-22 years)
only to the level displayed by intact males of
the same age (Chambers and Phoenix, 1983).
Increasing the daily dose of TP from 10 to 50 mg
does not produce a further increase in sexual
performance.
Not all males show the same rate or extent of
loss of sexual behavior in old age any more than
all young adult males show the same rate or
extent of decline in sexual behavior after
castration (Resko and Phoenix, 1972; Phoenix,
1973b; Phoenix et al., 1973). Baseline levels of
sexual behavior displayed after castration vary
widely among individuals, and serum T levels
required to reinstitute the full pattern of
sexual behavior may also be expected to vary
widely (Michael and Wilson, 1974). in this
study, we proposed to determine threshold
amounts of TP that would increase individual
components of sexual behavior in old, castrated
males to levels displayed by equally old, intact
males. We compared levels of serum T and DHT in
intact and castrated males before and after each
series of behavioral tests that corresponded to
each graded dose level of TP, and related
behavioral changes to changes in serum hormone
levels. [...]
Discussion : Testosterone failed to
induce mounting in two castrated males, and
three of the castrated monkeys (50%) did not
achieve intromission or ejaculation in 21 weeks
of TP treatment. In behavioral tests 14 years
earlier, the males had responded to T by
displaying the complete copulatory pattern; but
3 years ago, only two of the three males
ejaculated when injected with TP (10 mg/day),
and one male displayed only mounting behavior.
This male did not mount in the pair test
reported here but did masturbate to ejaculation
during two tests. Ejaculatory plugs were
frequently observed in the drop pans below the
home cage of this and the other males; however,
systematic records of ejaculatory plugs were not
kept during this study (but see Phoenix and
Jensen, 1973). Under very different
circumstances, adult rhesus males that sustained
lesions to the medial preoptic-anterior
hypothalamic area no longer copulated with
females but continued to masturbate to
ejaculation (Slimp et aL, 1978). The
behavioral deficits could not be attributed to a
T deficiency. Long-term castration and old age
(or old age alone) could lead to changes in
comparable areas of the central nervous system
with similar behavioral consequences. Whatever
the explanation for the loss of responsiveness
to T, the absence of copulatory behavior by
these males was not a function of low serum
levels of T. During the last series of
tests, the mean serum T level for castrated
males that did not intromit was 23.9 ng/ml; for
those that did, the mean was 24.5 ng/mI. The
mean T level for intact males in this test
series was 4.2 ng/ml. In previous studies the
mean levels of serum T at 0900 h in old males
were 3.9 and 6.2 ng/ml, and at 2 100 h the means
were 11.71 and 16.9 ng/ml (Chambers and Phoenix,
1981; Chambers et al., 1982). We have also shown
that sexual performance levels are no higher
when males are tested at 2 100 h when T levels
are high than when tested at 0900 h when levels
are low (Chambers et al., 1982). The fallure to
mount and intromit in the present study likewise
cannot be ascribed to a deficiency in serum DHT
levels.
In contrast to the three castrated monkeys
whose performance we just described, the other
three males mounted and achieved intromission
without T treatment. Their relatively high
levels of performance account for the absence of
statistically significant differences between
castrated and intact males for these behaviors
during the the first test series. After 1 week
of treatment with TP (0.004 mg/kg, or about
0.036 mg/male), one of the castrated monkeys
ejaculated; after 2 weeks of treatment, a second
male ejaculated. Their mean serum T level at the
end of the series was 0.39 ng/ml. The third
castrated male to ejaculate did so after the
first week of treatment with 0.016 mg of TP/kg
(or about 0.14 mg/male). His serum T level at
the end of the series was 1.91 ng/ml. Michael et
al. (1984) reported an increase in ejaculation
in recently castrated adult rhesus males each
given 0.05-0.10 mg of TP and whose plasma T
levels were between 1.20 and 2.05 ng/ml. In
spite of the many differences in procedures and
behavioral measures, the threshold values for
ejaculation that we found in old
long-term-castrated monkeys were very similar.
However, supraphysiologic levels of T did not
increase performance rates in responding males
and did not induce nonresponders to mourit,
intromit, and ejaculate.
Ejaculation was the only component of sexual
behavior to differentiate untreated, castrated
males from intact males. The percentage of tests
with ejaculations and the proportion of males
that ejaculated differed during the first test
series when vehicle alone was injected.
Treatment of the castrated males with 0.004 mg
of TP/kg during the second series of behavioral
tests eliminated all statistically significant
group differences in sexual performance. During
the fifth series of tests, however, there was a
statistically significant difference between the
groups in the proportion of males that
ejaculated. In that series, two of the six
castrated males ejaculated, the saine proportion
that ejaculated in the second and sixth series
of tests. But in the fifth series, all five of
the intact males ejaculated, and so there was a
statistically significant difference in
proportion.
Yawning
behavior is sexually dimorphic and
T-dependent (Phoenix et al., 1967; Phoenix and
Chambers, 1982c), but its role in the total
pattern of mating behavior is a matter of
conjecture. The proportion of intact males that
yawned during series 1 and 2 was greater than
that of castrated males, and the rate of yawning
was also higher in intact males during Series 1
and 3. Not until mean serum levels of T reached
5.66 ng/ml during the fourth series of tests
were all differences in yawning between the
groups eliminated.
There were no differences in any measure of
sexual behavior across the six treatment
conditions for the intact males (p>0.05); but
among the castrated males, contact, mount, and
yawn rates differed significantly overall. When
they received 0.064 mg/kg of TP, they contacted
females at a higher rate than when they received
0.004 and 0.016 mg of TP/kg, but the rate was no
greater than that shown when they were not
treated. Nor did the contact rate differ from
the rate in tests involving higher doses of T.
It is difficult to attach much meaning to this
difference or to that found in mounting rate
across treatments since the Newman-Keuls test
did not identify significar- differences in
mounting rate between any specific test series
despite a significant F value.
However, the yawning rate was significantly
higher in series 5 and 6 than in series 1-4. The
same was true for differences in serum T levels
among castrated males. The orderly increase in
yawning rate with increases in serum T levels
suggests a causal relationship between the two.
It is ironic that T, "the male sex hormone,"
is more closely associated with the yawning rate
than with the mounting or intromitting
rates.
The importance of female partners to the
outcome of pair tests of sexual behavior in
nonhuman primates is well known (Herbert, 1968;
Phoenix, 1973a; Goy, 1979; Michael et al.,
1984), and the partner is especially important
in tests involving old males. We have shown that
the performance levels of old males paired with
females chosen at random was significantly lower
than that of young and middle-aged males tested
with the same females; but, when old males were
paired with selected preferred females, their
performance level equalled that of young males
(Chambers and Phoenix, 1984). In the study
reported here, the percentage of tests in which
intact males ejaculated increased significantly
when these males were tested with our most
highly receptive and proceptive females. The
increase for castrated males, however, was not
significant. The three males that did not
intromit or ejaculate in earlier tests did not
do so with the highly proceptive females.
We considered the possibility that a
dichotomy existed in the population of old,
castrated males. There were old, castrated males
whose sexual behavior, especially ejaculation,
increased with T treatment and old males whose
level of sexual behavior could not be enhanced
by T. This dichotomy is in fact what we found in
the sample of animals from this study. A more
parsimonious explanation of our findings,
however, is that with age all males eventually
become unresponsive to T. The process is
gradual, and signs of decline are evident in the
reduced levels of sexual behavior observed in
intact males whose serum levels of T have not
declined. Some old, intact males stop
intromitting and ejaculating altogether.
Eventually, the level of sexual behavior of very
old, intact males should not differ from that of
old, castrated males. The hypothesis is
testable, but not easily so, because most males
die before reaching that age.
