I monitored 24-29 breeding groups of Harris' Hawks (Parabuteo unicinctus) in southeastern New Mexico during a 40-month period. Each year 29-70% of these groups made a second breeding attempt in either late summer or autumn, often following successful spring breeding. The interbrood period was positively correlated with the size of the first brood. Clutches were initiated between 17 February and 1 November. The mean number of hawks fledged from autumn nests was less (0.94) than that produced from spring (1.30) and summer (1.36) nests. The composition of prey remains found at nests in different seasons was similar. The number of second nests initiated was correlated with the number of lagomorphs but not with precipitation. Harris' Hawks did not time nesting to coincide with peak population levels of their principal prey; rather, this species has a flexible and prolonged breeding season that may have evolved in part to minimize the impact of stochastic climatic conditions that increase the probability of nest failure. Received 21 February 1986, accepted 27 August 1986.
Department of Biology, University of New Mexico, Albuquerque, New Mexico 87131 USA
AvI^N breeding seasons typically coincide
with circumstances such as abundant food, de-
creased competition, minimal predation, or be-
nign climatic conditions (ultimate factors) that
result in the successful recruitment of young
into the population (Immelmann 1971). Birds
breed in different months in different regions
of the world, but each species usually lays its
eggs at about the same date each year. The ini-
tiation of reproduction is triggered by the
proximate cues that most reliably predict the
onset of key ultimate factors (Perrins 1970, Im-
melmann 1971).
For most temperate-region birds, photoperi-
od is the most important proximate cue used to
time the breeding seasons (Murton and West-
wood 1977). In the tropics, where photoperiod
varies little, or in unpredictable desert envi-
ronments, precipitation is regarded as an im-
portant cue in the initiation of reproduction
(see Brown and Britton 1980). In both temper-
ate and tropical regions, food supply is proba-
bly the most important ultimate factor in the
determination of avian breeding seasons (Lack
1968, Perrins 1970, Immelmann 1971). At
northern latitudes, therefore, breeding seasons
typically coincide with the annual spring and
summer pulse of productivity. A few species
such as the Red Crossbill (Loxia curvirostra; Tor-
doff and Dawson 1965), Tricolored Blackbird
(Agelaius tricolor; Orians 1960, Payne 1969), and
Pinyon Jay (Gymnorhinus cyanocephalus; Ligon
1971) also breed in the autumn.
85
Birds of prey normally raise only one brood
in a year (Newton 1977, 1979), although some
of the smaller raptors, such as the American
Kestrel (Falco sparverius) and the Black-shoul-
dered Kite (Elanus caeruleus), occasionally rear
two broods in one year (Pickwell 1930, Stah-
lecker and Griese 1977, Toland 1985). Of the
larger species, the Common Barn-Owl (Tyto
alba) breeds almost year-round, primarily in
tropical environments (Otteni et al. 1972, Bunn
et al. 1982, Lenton 1984). De Vries (1975) re-
corded three successful second-brood attempts
by the equatorial Galapagos Hawk (Buteo gala-
pagoensis). Harris' Hawks (Parabuteo unicinctus)
in temperate North America are known to nest
in the autumn in Texas (LeSassier and Williams
1959, Brannon 1980), Arizona (Radke and Kli-
mosewski 1977, Ellis and Whaley 1979), and
New Mexico (Pache 1974). Mader (1975) and
Whaley (1986) documented that Harris' Hawks
laid eggs as early as January and fledged young
as late as October. In addition, Mader (1975)
observed five cases in which breeding groups
with one or more banded adults renested after
successfully rearing a first brood. Harris' Hawks,
therefore, are unusual among large Temperate
Zone raptors in having a prolonged breeding
season and in rearing successive broods in one
year.
Here I report year-round observations on the
reproductive efforts of breeding groups of Har-
ris' Hawks in a New Mexico population over a
40-month period, and describe the temporal
patterns of breeding and successive nesting at-
tempts. I then relate these patterns to fluctua-
tions in available prey and precipitation to
identify potential proximate and ultimate fac-
tors.
STUDY AREA AND METHODS
The study area was approximately 150 km 2 and was
located in Eddy and Lea counties in southeastern New
Mexico. The habitat is mesquite-oak shrubland with
rolling sandy soils. The dominant shrubs include
mesquite (Prosopis glandulosa), Havard's oak (Quercus
havardii), sand sage (Artemisia filifolia), snakeweed
(Gutierrezia sarothrae), yucca (Yucca cornpestris), and
creosotebush (Larrea tridentata). Harris' Hawks nested
primarily in larger mesquite and soapberry trees (Sa-
pindus drummondi) dispersed throughout the study
area. This region of New Mexico is dry with a normal
annual total precipitation of 30.25 cm (30-yr average;
NOAA 1973). Typically, there is a daily temperature
fluctuation of 15øC or more. Summer temperatures
can exceed 40øC during the day and may drop below
20øC at night. In winter, night temperatures below
- 10C are recorded regularly. Snow or ice storms can
occur in the study area from mid-October through
April, and violent summer rainstorms are common.
Data were collected between August 1980 and No-
vember 1983. I made at least one visit to the study
area in each month except December and February.
Each year (I 981-1983) all known nesting ranges were
searched thoroughly at least three times (in spring,
summer, and autumn) for the presence of active nests.
If no nests were found during any of these searches,
I reexamined ranges without active nests until I was
reasonably sure that no nesting attempt had been
missed. If a nest failed early in the incubation stage,
this often could be determined by the presence of
broken eggshells either in the nest or on the ground.
Given the intensity and frequency of my search ef-
forts, conspicuousness of nests, and scarcity of suit-
able nest trees, I believe I missed very few breeding
attempts that failed during early incubation. Fledg-
ing was recorded as successful only after volant young
were seen away from the nest. The nesting area of
Harris' Hawks in New Mexico does not seem to be
defended actively (Bednarz 1986) and thus should
not be referred to as a territory. Therefore, I use the
term "nesting range," defined as the area habitually
occupied by a breeding group of hawks and within
which they placed their nests.
Adults and nestlings were banded with unique
color combinations of vinyl tarsal bands. Bands were
made with Norcross virgin vinyl (Norcross Indus-
tries, West Palm Beach, Florida) strips (7 cm in length
for males and 7.5 cm for females; width = 1.6 cm)
riveted together, leaving a trailing tab approximately
2.5 cm in length. Adults were trapped with bal-chatri
(Berger and Mueller 1959) and padded steel leg traps.
During the study, I color-marked 163 hawks. By No-
vember 1983 about half of the approximately 75 adult-
plumaged hawks in the population were banded.
Age of nestling hawks was estimated by use of the
regression equation: estimated age = 8.73991 +
(length of seventh primary in mm x 0.19797), deter-
mined from 200 measurements of 24 known-age
nestlings of both sexes (F = 4,135, P < 0.0001, r 2 =
0.95). Males and females exhibited no significant dif-
ferences in the growth of this primary feather during
the measurement period. The relationship of sev-
enth-primary length and age is linear after the feath-
er begins emerging at about 8 days of age (Fig. I).
All nestlings were measured and weighed during
each nest visit, and hatch dates were estimated with
the above equation. Because most nests contained
multiple nestlings that were measured several times,
I was able to calculate an average estimated date of
hatching. I estimated dates for clutch initiation and
fledging for 78 nests by using standard incubation
(34 days, n = 4 clutches) and brood-rearing periods
(46 days, n = 12 nestlings). Males tended to fledge
earlier than females (mean = 43.5 days, SD = 1.76,
n = 6 vs. mean = 48.6 days, SD = 3.14, n = 6; P >
0.I).
For nests that failed before the eggs hatched (n =
33 nests), the date of nest initiation was estimated
based on the length of time the nest was known to
be active and when the adults abandoned the nest.
For these nests the estimated date of nest initiation
could be in error by as much as 14 days. Clutch ini-
tiation for 4 nests that were found just after the young
fledged was determined by using the estimated date
of fledging and subtracting 80 days (34-day incu-
bation + 46-day brood-rearing periods).
Food consumed by nestling Harris' Hawks was de-
termined by identifying bones, bone fragments,
feathers, and other prey parts found in and beneath
nests, either singly or within pellets. For identifica-
tion, I compared remains with known reference spec-
imens at the University of New Mexico Museum of
Southwestern Biology. Mammal fur was not used in
this analysis because it was difficult to identify con-
clusively, and preliminary analysis of fur paralleled
the results obtained from skeletal remains. The min-
imum number of individuals of each prey taxa pres-
ent at each nest was the maximum number of a spe-
cific bone for each species represented in the remains.
As I found no reliable character to distingiush cot-
tontail rabbit (Sylvilagus audubonii) from immature
black-tailed jackrabbit (Lepus californicus) limb bones,
I recorded only bones larger than those of adult cot-
tontails as jackrabbits. Biomass of the prey consumed
was estimated by multiplying numbers of prey taken
by the average wet mass of prey types, assuming no
waste. Mean mass values for three dominant prey
types, desert cottontail, woodrat (Neotoma spp.), and
spotted ground sqtlirrel (Sperrnophilus spilosorna), were
based on samples collected at the study area (561 g,
SD = 197, n = 17; 217 g, SD = 29, n = 288; and 117
g, SD = 73, n = 32, respectively). The intermediate so-
or overall mean mass values reported by Steenhof
(1983) were used as biomass estimates for rarer prey. 4o-
Road counts of cottontails and black-tailed jackrab-
bits were conducted twice each month from May 1981
(3 30'
to October 1983, except February and December. The
census was run in alternating directions over the same z_
16-km route beginning 15-25 min before sunset and 2o-
ending at sunset. The transect route dissected almost
the entire study area (ca. 150 km2), and I feel that the o-
mean values for each month were representative of
relative numbers of lagomorphs available to the
hawks.
For several analyses, nests were classified as spring,
summer, or autumn nests: spring nests = young
hatched before 1 June, summer nests = young hatched Fig.
between 1 June and 31 August, and autumn nests =
young hatched 1 September or after. Standard para-
metric analyses were used where suitable; however,
most data were analyzed with Kruskal-Wallis or
Spearman's correlation statistics because distribu-
tions deviated from normal (Sokal and Rohlf 1981).
RESULTS
Temporal breeding patterns.--In autumn 1980,
I found 7 active Harris' Hawk nests, 2 initiated
in August and 5 initiated in September, in an
area about one-third the size of the study area.
Based on findings in subsequent years, I be-
lieve I thoroughly searched 10 Harris' Hawk
nesting ranges at that time. If all of these ranges
were occupied in 1980, then 70% of this sample
of hawk ranges were used for autumn breeding
in that year. Of the 7 nests, 5 failed during rain
(1) or snowstorms (4) and 2 were successful.
In 1981, I thoroughly searched 22 nesting
ranges throughout the year. Range 21 was not
checked adequately for breeding in the au-
tumn, and range 25 was discovered in late Au-
gust and thus was not examined during the
previous spring (Fig. 2). The earliest clutch in
1981 was initiated on about 17 February. By
May, 20 groups (91% of those monitored year-
round) were tending active nests in stages
varying from egg laying to fledging (Fig. 2). Of
the 16 groups that successfully fledged young
during their spring breeding attempt, 8 (50%)
nested a second time and 6 of these (37.5%)
successfully produced a second brood. Breed-
ing pair 7 initiated its second clutch only 9 days
after its first young fledged, and then laid a
third clutch of eggs in October after success-
fully fledging 2 broods of one young (Fig. 2).
0 4'0 8'0 ' 1'0 10 20
LENGTH OF SEVENTH PRIMARY IN MM
1. Regression of nestling age on seventh pri-
mary length.
Three of the 5 breeding units that were unsuc-
cessful in their first attempt successfully reared
young from a second nest. Group 9 made four
nesting attempts, and the female laid a mini-
mum of 14 eggs in 4 clutches in 1981 (Fig. 2).
Group 4 initiated its first and, as far as known,
only breeding attempt in September. The latest
clutch of eggs was laid between 16 October and
1 November by group 9, but failed before the
eggs hatched. Had it been successful, young
would have fledged in January. The latest suc-
cessful breeding effort recorded (group 4) was
initiated on 16 September, and one young left
the nest about 5 December and was seen alive
in January. Group 19 was caring for nestlings
until at least the second week in December,
when the nest was destroyed by an unknown
predator. These observations suggest that un-
der ideal conditions Harris' Hawks in southern
New Mexico can fledge young well into De-
cember. In 1981, active hawk nests were pres-
ent on the study area for more than 300 days
of the year.
In 1982 initial breeding attempts were much
more synchronous than in 1981 (Fig. 3). Of the
28 breeding groups followed year-round, 93%
were nesting by early April. Seven (33%) of the
21 groups that were successful during their first
attempt bred a second time, and 5 of these
fledged two broods. Three of the 6 groups
whose first nests failed tried a second time; all
of these were unsuccessful. Group 9 initiated
its first and probably only breeding effort of
the year in September. The earliest 1982 clutch
(group 17) was laid on 3 March, and the last
2O÷
z 16
uJ
x 1981
G
d F M A M d d A $ O N D
Fig. 2. Temporal pattern of Harris' Hawk breeding in 1981. Histogram shows the number of groups
actively tending nests within 20-day intervals. The numbers to the right of the horizontal bars indicate the
number of young fledged, and X's indicate nest failure.
young were fledged on about 17 November
(group 7).
The frequency of groups breeding through-
out the year exhibited a distinctly bimodal pat-
tern. This probably was due to the interbrood
period (days from fledging of the first brood
to the initiation of the second clutch), which
was significantly greater (P = 0.006) in 1982
(mean = 81.0 days, SD = 27.1, n = 7) than in 1981
(mean = 40.8 days, SD = 20.6, n = 8), when some
1982
I////// Ix [T/ZI = INCUBATION
V////IA la [ = BROOD-REARING
d F M A M d d A $ O N D
Fig. 3. Temporal pattern of Harris' Hawk breeding in 1982. See Fig. 2 for further details.
Fig. 4.
6
.... YA Y/? "
Vll///// I: [7773 = INCUBATION
J F M A M J J A S O N D
Temporal pattern of Harris' Hawk breeding in 1983. See Fig. 2 for further details.
successful groups renested almost immediately
(e.g. group 7).
First breeding attempts in 1983 were rela-
tively asynchronous and autumnal nests were
few (Fig. 4). The greatest frequency (93%) of
active nesting by the 29 monitored groups oc-
curred during the second week of April. Of the
17 groups successful in the spring, only 1 (5.9%)
nested a second time. Seven (63%) of the 11
groups that were initially unsuccessful at-
tempted a second nest; 6 of these groups even-
tually were successful. The female of pair 11
laid eggs for the first time in 1983 on 30 Au-
gust. The latest f/edging date recorded in 1983
was 8 November (group 12).
Nest success by season.--There were no signif-
icant trends (Kruskal-Wallis tests, P > 0.1) in
either clutch size or number of young fledged
among the three seasons during the study pe-
riod (Tables 1 and 2). The number of young
fledged per nest in autumn tended to be low
in all years except 1982. The percentage success
of autumn nests for 1981-1983 (44.4%, 62.5%,
and 50.0%, respectively) was consistently be-
low that of spring nests (78.6%, 76.9%, and
55.6%, respectively). Likewise, nest success in
autumn 1980 was very low (28.6%). The per-
centage success (X 2 = 4.3, P = 0.12, df = 2) and
the number of young fledged per successful nest
(Kruskal-Wallis test, P > 0.1) were not signifi-
cantly different among seasons (Table 2).
Interbrood period.--In 1981 there was a signif-
icant linear relationship between the number
of young fledged and the length of the inter-
brood period (F = 13.2, P = 0.01, r 2 = 0.69; Fig.
5). The trend in 1982 was similar, but did not
approach significance (F = 0.7, P = 0.45, r 2 =
0.12; Fig. 5). In 1983 only one group renested
after successfully rearing a brood.
Diet.--Harris' Hawks relied heavily on cot-
TABLE 1. Mean dutch size for Harris' Hawks in different seasons.
Spring Summer Autumn
Mean SD n Mean SD n Mean SD n
1980 No data No data 2.50 0.71 (2)
1981 3.50 0.55 (6) 2.80 0.94 (15) 3.17 0.41 (6)
1982 3.05 0.81 (21) 2.67 0.58 (3) 3.50 0.71 (2)
1983 2.90 0.54 (21) 3.25 0.50 (4) No data
1981-1983 3.04 0.68 (48) 2.86 0.83 (22) 3.25 0.46 (8)
TABLE 2. The mean number of Harris' Hawk young fledged per nest attempt and successful nest in different
seasons.
Spring Summer Autumn
Mean SD n Mean SD n Mean SD n
1980 No data No data 0.71 1.25 (7)
1981 1.43 1.09 (14) 1.38 1.20 (16) 0.50 0.76 (8)
1982 1.46 1.03 (26) 0.75 1.50 (4) 1.43 1.40 (7)
1983 1.07 1.17 (27) 1.60 1.17 (10) 1.00 1.41 (2)
All nest attempts
1981-1983 1.30 1.10 (67) 1.36 1.22 (30) 0.94 1.14 (17)
Successful nests
1981-1983 1.89 0.80 (46) 2.05 0.89 (20) 2.10 0.74 (10)
tontails for food throughout the year (Table 3).
Cottontails (or young jackrabbits no larger than
adult cottontails) comprised over 80% of the to-
tal prey biomass in each year and in each sea-
son. Larger jackrabbits were most important in
late spring, although they were taken occasion-
ally throughout the year. Jackrabbits are un-
doubtedly more important in the Harris'
Hawk's diet than Table 3 indicates, because I
could not distinguish bones of juvenile Lepus
from those of juvenile Sylvilagus. Woodrats
(Neotoma spp.) also were an important prey, es-
pecially during spring. Quail and a variety of
other birds were relatively common in the diet
during summer and autumn (Table 3). Insects
were eaten by nestlings in the summer, but
these contributed little to the diet (Table 3).
Most of these insects (68.6%) were small cole-
opterans.
Diets in spring and autumn were similar in
that lagomorphs and other mammalian prey
predominated (Table 3). There were no differ-
ences between years (P > 0.1); however, a sig-
nificant trend was found when comparing the
number of mammal vs. other prey types taken
among seasons (X 2 = 51.3, P < 0.0001, df = 2).
This was due primarily to an increase in con-
sumption of insects and reptiles in the summer
(Table 3). Nevertheless, mammals made up
94.5% of the biomass of the food consumed by
nestlings during the summer because of the
great size of lagomorphs as compared with oth-
er prey types. Some bias in the prey analysis
undoubtedly exists, in that smaller items are
more easily distorted or fragmented beyond
recognition and also are more difficult to find
than larger bones. However, I believe this bias
was not great because (1) complete limb bones
and skulls of medium-size rodents and reptiles
were common in the remains, (2) data from
Swainson's Hawk (Buteo swainsoni) nests in the
same area reveal that smaller types of prey were
found in significantly (P < 0.0001) greater pro-
portions (Bednarz 1986), and (3) some ex-
tremely small and rare animals on the study
site were found among the prey remains (e.g.
Reithrodontomys sp.) (Gennaro 1982).
Lagomorph numbers, precipitation, and nest ini-
tiation.--To examine the role of potential prox-
imate factors involved in the timing of Harris'
Hawk breeding, I used Spearman's correlation
analysis (Sokal and Rohlf 1981) to identify pat-
terns among 2-month precipitation (total pre-
cipitation during month of nest initiation and
previous month), mean monthly cottontail
numbers, mean total Iagomorph numbers (cot-
tontails and jackrabbits), the number of nests
initiated in each month, and the number of sec-
ond nesting attempts initiated each month. The
observation period extended from May through
October in 1981 (no Iagomorph data were col-
lected during March and April), and March
through October in 1982 and in 1983. For this
analysis, I included second nesting efforts after
both successful and unsuccessful first-breeding
attempts. I believe that the propensity for a
particular group to renest may be best mea-
sured in this manner, given that in some years
certain groups do not renest if the first brood
is lost. In addition, second nesting attempts af-
ter nest failure cannot be equated with the
clutch replacement observed in other large
raptors, because Harris' Hawks may initiate
second nests if broods about to fledge are lost
and may lay second clutches up to 6 months
after failure of the first clutch of the year.
No patterns were detected between the num-
ber of Harris' Hawk nests initiated and cotton-
60
2
"" I 2 3 4
uJ
Q.
0 ß
0 oo- 1982
UJ 80-
;.- ß
60
40-
20-
NUMBER FLEDGED IN 8ROOD
Fig. 5. Relationship of the number of Harris'
Hawk young fledged in the first brood to interbrood
period in days for 1981 and 1982.
tails observed (r = 0. I, P = 0.67, n = 22 months),
total lagomorphs (r = 0.I, P = 0.67, n = 22), or
2-month precipitation (r = -0.35, P = 0.11, n =
22) when data for all years were combined.
However, number of second nests initiated cor-
related significantly with the numbers of cot-
tontails (r = 0.53, P = 0.01, n = 22) and of total
lagomorphs (r = 0.52, P = 0.01, n = 22). The
relationships between Iagomorph numbers and
renesting were consistent for analyses done
within years, but were not statistically signifi-
cant. In 1982, the number of second nesting
attempts was almost significantly correlated
with cottontail numbers and total lagomorphs
(r = 0.63, P = 0.09, n = 8, and r = 0.63, P = 0.07,
n = 8, respectively). Correlations of other vari-
ables changed signs between years and, except
for one case, did not approach significance. The
exception is that 2-month precipitation was
negatively correlated with the number of all
nests initiated in 1983 (r = -0.78, P = 0.02, n =
8). This trend was not seen in 1981 and 1982,
suggesting that the 1983 pattern may be spu-
rious.
Analyses including only nests initiated after
the first brood fledged successfully likewise re-
vealed that the number of second nests initi-
ated was significantly correlated with the num-
ber of cottontails (r = 0.78, P < 0.01, n = 12)
and total lagomorphs (r = 0.73, P < 0.01, n =
12). Only the months May through October,
when the initiation of such nesting attempts
was observed, were included in this analysis. I
omitted 1983 because only one second nesting
occurred. Again, 2-month precipitation was not
correlated with the number of second nests ini-
tiated (P > 0.2).
A final analysis was done to examine the sug-
gestion that hawks may time breeding so that
fledging coincides with the peak period of prey
availability (see Immelmann 1971, Snyder and
Wiley 1976). No relationship was found be-
tween the number of nests that fledged young
and cottontail numbers (r = 0.2, P = 0.4, n =
20 months) or total lagomorphs (r = 0.31, P =
0.2, n = 20). The months May through Novem-
ber were included in this analysis because Iago-
TABLE 3. Percentage of prey types represented in food remains found in and near nests of Harris' Hawks
breeding in different seasons.
Spring (n = 422) Summer (n = 237) Autumn (n = 230)
Prey type Frequency Biomass Frequency Biomass Frequency Biomass
Sylvilagus auduboni 65.4 84.8 46.0 88.2 70.0 90.6
Lepus californicus 2.8 6.6 0.4 1.4 0.4 1.0
Neotoma spp. 8.3 4.2 1.7 1.3 3.9 2.0
Spermophilus spp. 5.0 1.3 5.9 2.4 1.7 0.5
Other rodents 2.1 0.3 5.5 1.2 3.9 0.6
QuaiV 3.8 1.6 5.1 3.1 9.6 3.9
Other birds 2.6 0.9 5.1 2.1 4.3 1.4
Reptiles 4.7 0.2 5.1 0.4 0.9 Trace
Insects 5.2 Trace 25.3 Trace 5.2 Trace
Colinus virginianus and Callipepla squamata occur on the study site.
3O'
1981
/TOTAL LAGOMORPHS
COTTONTAILS
M A J J S O N
n 15
o
0
z
1982
TOTAL LAGOPH
J F M A M J J
A S O N
1983
15 TOTAL LA 6
,o
J F M A M J J A S 0
MONTH
z
Fig. 6. Mean numbers of cottontails and total
lagomorphs counted on census route and numbers of
second nests initiated (histograms) each month in
1981-1983.
morph data were available and 98.7% of the
successful fledgings occurred in this period.
Lagomorph numbers peaked primarily in
mid- to late summer, which corresponded with
the period during which most second nests were
initiated. Although iagomorph numbers were
high throughout 1981, 50% of the renestings
were initiated in July and August when prey
populations peaked (Fig. 6). The relationship
was similar in 1982, when 70% of the second
nests were initiated during the July-August ia-
gomorph population peak (Fig. 6). Reproduc-
tive success was low in 1983, and all but one
second breeding attempt were initiated after
nest failures. These nest starts corresponded with
the late spring increase in iagomorph popula-
tions (Fig. 6).
The cottontail population was high in 1981
and apparently crashed during the winter of
1981-1982 (Table 4). Jackrabbit numbers, on the
other hand, showed no significant trends dur-
ing the three years. The greatest tendency to
renest occurred in 1981, when cottontails were
most numerous. In that year, 55% of the 20
spring or summer breeding attempts were fol-
lowed by a second effort (Fig. 2). In addition,
the interbrood period was significantly shorter
in 1981 (mean = 40.8 days) compared with 1982
(mean = 81.0 days). In contrast, in 1982 and
1983, both relatively poor cottontail years, the
percentages of spring and summer breeding at-
tempts followed by a second nesting were 37.1%
and 28.6%, respectively (Figs. 3 and 4).
The incidence of groups nesting in autumn
1980 was high (7 of 10 known breeding ranges
contained active nests). This corresponded to a
peak in rabbit populations, as deduced from
Iagomorph counts conducted by Gennaro (1982)
on the study area in 1977-1979 and 1981. Dur-
ing the 4 yr I monitored the Harris' Hawk pop-
ulation, the frequency of autumn nesting and
second breeding attempts was greatest during
the 1980-1981 period of cottontail abundance.
In 1982 and 1983 the decrease of late nesting
apparently paralleled a decline in iagomorph
populations (Table 4).
DISCUSSION
Harris' Hawks in southeastern New Mexico
commonly breed in the spring, summer, and
autumn when environmental conditions are
suitable. The breeding season normally com-
mences about mid-February and may continue
into November or early December. Most breed-
ing attempts in late summer and autumn are
undertaken by groups that bred in the previous
spring. During a good prey year (1981) the per-
centage of groups nesting in the fall was more
than 50% and in 1980 it might have been as
high as 70%. In Texas, Brannon (1980) found
Harris' Hawks nesting from February through
November. Mader (1978) and Whaley (1986) re-
ported that the breeding season of Harris'
Hawks in Arizona ended in late October, al-
though Whaley (1979) noted one exception.
These studies, as well as more recent work (J.
Dawson pers. comm.), suggest that the fre-
quency of autumnal nesting in Arizona is much
lower than I recorded in New Mexico.
In species that breed successively (mostly
passerines), it has been predicted that if food
conditions are poor or if broods are large, one
should expect longer intervals between broods,
as energy and protein reserves will be depleted
by the care of early broods (Ricklefs 1974, Jones
and Ward 1976). In 1981 ! found a significant
positive relationship between the size of the
first brood and the interbrood period, support-
ing this suggestion (also see Murphy 1978,
Smith and Roff 1980, McGillivray 1983). This
trend also was present in 1982, but was not
statistically significant. Moreover, the inter-
brood period was significantly shorter in 1981,
the best prey year, than in 1982, a relatively
poor prey year. These results support the view
that available food resources are an important
factor in an individual's ability to initiate a sec-
ond clutch. Second nesting attempts by Harris'
Hawks did not result in smaller clutches or
broods of young (Table 1), as McGillivray (1983)
found for the House Sparrow (Passer domesti-
cus). These results do not support Perrins's
(1970) suggestion that late clutches should be
small so that incubation can start earlier and so
that each young will be better nourished.
Commencement of breeding in late winter
and early spring was not correlated with either
precipitation or iagomorph abundance. Both the
numbers of cottontails and precipitation were
low in March, making it highly unlikely that
these were proximate cues that triggered the
onset of spring breeding. Variation in day
length is well substantiated as a proximate tim-
ing factor for many species, whereas there is
little evidence to suggest that increasing tem-
peratures trigger breeding (Lofts and Murton
1968). Thus, Harris' Hawks in New Mexico
probably are similar to Temperate Zone species
in that they use photoperiod to "predict" the
impending increase in prey numbers.
Harris' Hawk is primarily neotropical in dis-
tribution (Brown and Amadon 1968), and may
have only recently extended its range into the
southwestern United States. If this is the case,
I suggest that the "ancestral" tropical and sub-
TABLE 4. Mean numbers of cottontails and jackrab-
bits recorded on all road counts done between May
and September in different years. Different letters
indicate a significant difference between years at
the P < 0.05 level (analysis of variance).
1981 1982 1983
(n = 9) (n = I0) (n = I0)
Mean SD Mean SD Mean SD
Cottontails 14.3 10.9 3.0 2.3 3.6 !.8
Significance
group A B B
Jackrabbits !0.3 2.9 !0.4 5.7 !3.0 4.6
Significance
group A A A
tropical Harris' Hawks bred in response to re-
source availability, as many tropical species do
today. Whether tropical Harris' Hawks have or
had a defined breeding season or were contin-
uous breeders probably depended on local fluc-
tuations in prey populations. As the hawk
moved into more seasonal temperate climates
and was subjected to a more or less reliable
burst of prey in spring, they may have become
more sensitive to increasing day lengths to take
maximal advantage of the spring increase in
lagomorphs. Following the development of a
sensitivity to photoperiod, the actual time an
individual female hawk laid eggs would de-
pend on the availability of sufficient resources
to compensate for the energy and protein de-
mands of egg formation (Newton 1979). This
proposed sequence is similar to that suggested
by Lofts and Murton (1968) for the evolution
of a photoperiodically controlled cycle in Tem-
perate Zone species derived from equatorial
birds not dependent on photostimulation. Be-
cause Harris' Hawks in New Mexico can breed
throughout most of the year, it is unlikely that
they have the well-developed refractory period
found in many Temperate Zone species (Mur-
ton and Westwood 1977). Comparisons of
breeding season and ecological data between
temperate and tropical populations of Harris'
Hawks would be useful in evaluating these
ideas, but data on tropical populations do not
exist.
The abundance of cottontails in late summer
probably is the most critical factor responsible
for autumnal breeding attempts in Harris'
Hawks. Correlations between the number of
second nests initiated and the numbers of
available lagomorphs, as well as the decreased
interbrood period in the resource-abundant
year of 1981, support this contention. The im-
portance of suitable food resources as a proxi-
mate cue for other autumnal breeders has been
shown for the Tricolored Blackbird (Payne 1969)
and the Pinyon Jay (Ligon 1974). De Vries (1975)
noted that heavy rainfall preceded peak egg
laying in the Galapagos Hawk by 3 months and
speculated that responses of prey to precipita-
tion were probably the most important factor
triggering nesting.
The ultimate value of successive nesting and
autumn breeding must be that some late nests
produce birds with some probability of estab-
lishing themselves in the breeding population.
Late nests fail more often than earlier nests, but
this trend was not significant statistically. The
late summer and autumn nests were initiated
when available food resources were most abun-
dant, but the peak food demands of large nest-
lings occurred in November when cottontail
populations were declining. Obviously, autum-
nal breeding attempts are not timed so that
young will fledge during optimal conditions
(see Immelmann 1971). Rather, Harris' Hawk's
strategy apparently is to prolong the breeding
season when adequate food is available.
After fledging, Harris' Hawk young remain
on their natal range and share prey with their
parents for at least 3-6 months (Bednarz 1986).
Thus, parents presumably can assist offspring
from a late breeding attempt with food pro-
curement during the lean winter period. I sus-
pect that the extended parental care strategy
initially evolved in response to other factors
(e.g. those favoring the development of coop-
erative breeding; Bednarz 1986), rather than to
the late breeding habit.
Other factors of possible importance to Har-
ris' Hawk are unpredictable climatic events and
possibly predation that result in nest failure.
Although stochastic processes leading to nest
failure operate during all seasons, violent, un-
predictable summer storms could be particular-
ly important to Harris' Hawk breeding. These
storms may lessen the advantage of summer
breeding by increasing the probability of nest
failure at the time when available food re-
sources are at their yearly maxima. On the oth-
er hand, autumn, which often is climatically
mild, in some years might provide favorable
conditions for the production of young despite
the decreased density of cottontails. Two pat-
terns observed during this study are consistent
with this suggestion. First, precipitation in
summer occurs primarily in the form of violent
convectional thunderstorms that typically are
accompanied by strong gusting winds (Hough-
ton 1959), whereas precipitation in other sea-
sons generally does not involve destructive
winds. Second, the hawk population exhibits a
slight depression in breeding activity in mid-
summer (Figs. 2-4) even though the iago-
morph population is at its yearly peak (Fig. 6).
Whether this is the result of a recovery period
after the first nesting attempt of the year, the
need to provide parental care for the first brood,
or a postponement until arrival of the milder
conditions of autumn is unknown.
I suggest that the relatively high probability
of stochastic events (e.g. violent storms and
possibly predation) that result in nest failure
throughout the year, in combination with the
extended availability of an adequate food sup-
ply, has resulted in the prolonged and flexible
breeding season of Harris' Hawk. In addition,
the frequent violent summer storms may tend
to negate any advantage of breeding in this
season even though prey populations are at a
peak. These factors may have instead favored
the bimodal breeding pattern described here.
ACKNOWLEDGMENTS
I sincerely thank Joey Haydock, Steve Zack, J. Da-
vid Ligon, Greg Jillson, Jim Stuart, and several other
students for their assistance in the field. J. David Li-
gon offered suggestions and guidance throughout this
study. Beth Dennis produced most of the figures. Ed
Bedrick, Michael Collopy, John Faaborg, Paul Ma-
tray, Brian Millsap, Manuel Molles, George Anne
Thibodeau, and John Wiens provided comments at
various times during the project. The field work was
funded partially by the Department of Energy (Waste
Isolation Pilot Plant studies), the Frank M. Chapman
Memorial Fund, the University of New Mexico
(UNM) Graduate Student Research Allocation Com-
mittee, and the Graduate Research Allocation Com-
mittee of the Biology Department, UNM. This paper
is contribution No. 1 of the New Mexico Cooperative
Raptor Research and Management Program.
LITERATURE CITED
BEDNARZ, J. C. 1986. The behavioral ecology of the
cooperatively breeding Harris' Hawk in south-
eastern New Mexico. Ph.D. dissertation, Albu-
querque, Univ. New Mexico.
BERGER, D. D., & H. C. MUELLER. 1959. The bal-cha-
tri: a trap for the birds of prey. Bird-Banding 30:
18-26.
BRANNON, J. D. 1980ß The reproductive ecology of
a Texas Harris' Hawk (Parabuteo unicinctus harrisi)
population. M.S. thesis, Austin, Univ. Texas.
BROWN, L., & D. AMADON. 1968. Eagles, hawks and
falcons of the world, vol. 2. Great Britain, Coun-
try Life Books.
-, & P. L. BRITTON. 1980. The breeding seasons
of East African birds. Nairobi, East Africa Natu-
ral History Soc.
BUNN, D. S., A. B. WARBURTON, & R. D. S. WILSON.
1982. The Barn Owl. Calton, England, T. & A.
D. Poyser.
DE VRIES, T. J. 1975. The breeding biology of the
Galapagos Hawk, Buteo galapagoensis. Gerfaut 65:
29-57.
ELIas, D. H., & W. H. WHALEY. 1979. Two winter
breeding records for the Harris' Hawk. Auk 96:
413.
GENNARO, A. L. 1982. Ecology of amphibians, rep-
tiles, and mammals at the Los Medafios Waste
Isolation Pilot Plant (WIPP) project area of New
Mexico. Pp. 6.1-6.41 in Ecosystem studies at the
Los Medafios site, Eddy County, New Mexico,
vol. 2. Albuquerque, New Mexico, U.S. Dept.
Energy.
HOUGHTON, F.E. 1959. Climate of New Mexico. Cli-
matography of the United States No. 60-29. Sil-
ver Spring, Maryland, U.S. Dept. Commerce, Na-
tional Oceanic and Atmospheric Administration,
Environmental Data Serv.
IMMELMANN, K. 1971. Ecological aspects of periodic
reproduction. Pp. 341-389 in Avian biology, vol.
1 (D. S. Farner and J. R. King, Eds.). New York,
Academic Press.
JONES, P. J., & P. WARD. 1976. The level of reserve
protein as the proximate factor controlling the
timing of breeding and clutch-size in the Red-
billed Quelea Quelea quelea. Ibis 118: 547-574.
LACK, D. 1968. Ecological adaptations for breeding
birds. London, Methuen.
LENTON, G. M. 1984. The feeding and breeding
ecology of Barn Owls Tyto alba in peninsular
Malaysia. Ibis 126: 551-575.
LESASSIER, A. L., & F. C. WILLIAMS. 1959. Notes on
a late nesting of Harris' Hawks near Midland,
Texas. Wilson Bull. 71: 386-387.
LIGON, J. D. 1971. Late summer-autumnal breeding
of the Pition Jay in New Mexico. Condor 73: 147-
153.
1974. Green cones of the pition pine stim-
ulate late summer breeding in the Pition Jay. Na-
ture 250: 80-82.
LoFts, B., & R. K. MURTON. 1968. Photoperiodic and
physiological adaptations regulating avian
breeding cycles and their ecological significance.
J. Zool. London 155: 327-394.
MADER, W.J. 1975. Biology of the Harris' Hawk in
southern Arizona. Living Bird 14: 59-85.
ß 1978. A comparative nesting study of Red-
tailed Hawks and Harris' Hawks in southern Ar-
izona. Auk 95: 327-337.
McGIIZIVRY, W. B. 1983. Intraseasonal reproduc-
tive costs for the House Sparrow (Passer domes-
ticus). Auk 100: 25-32.
MURPHY, E. C. 1978. Seasonal variation in repro-
ductive output of House Sparrows: the deter-
mination of clutch size. Ecology 59: 1189-1199.
MURTON, R. K., & N.J. WESTWOOD. 1977. Avian
breeding cycles. Oxford, Clarendon Press.
NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRA-
TION. 1973. Monthly normals of temperature,
precipitation, heating and cooling degree days
1941-70 (New Mexico). Asheville, North Caro-
lina, U.S. Dept. Commerce, NOAA, Environ. Data
Serv., Natl. Climatic Center 81: 1-9.
NEWTON, I. 1977. Breeding strategies in birds of prey.
Living Bird 16: 51-82.
ß 1979. Population ecology of raptors. Ver-
million, South Dakota, Buteo Books.
ORIANS, G. H. 1960. Autumnal breeding in the Tri-
colored Blackbird. Auk 77: 379-398.
OTTENI, L. C., E.G. BOEN, & C. COTTAM. 1972. Pred-
ator-prey relationships and reproduction of the
Barn Owl in southern Texas. Wilson Bull. 84: 434-
448.
PACHE, P. H. 1974. Notes on prey and reproductive
biology of Harris' Hawk in southeastern New
Mexico. Wilson Bull. 86: 72-74.
PAYNE, R.B. 1969. Breeding seasons and reproduc-
tive physiology of Tricolored Blackbirds and
Redwinged Blackbirds. Univ. California Publ.
Zool. 90: 1-137.
PERRINS, C.M. 1970. The timing of birds' breeding
seasons. Ibis 112: 242-255.
PICKWEI, I, G. 1930. The White-tailed Kite. Condor
32: 221-239.
RADKE, E. L., & J. KIIMOSEWSK. 1977. Late fledging
date for Harris' Hawk. Wilson Bull. 89: 470-471.
RICKLEFS, R. E. 1974. Energetics of reproduction in
birds. Pp. 152-292 in Avian energetics (R. A.
Paynter, Jr., Ed.). Publ. Nuttall Ornithol. Club
No. 15.
SMITH, J. N.M., & D. A. ROFF. 1980. Temporal spac-
ing of broods, brood size, and parental care in
Song Sparrows (Melospiza melodia). Can. J. Zool.
58: 1007-1015.
SNYDER, N. F. R., & J. W. WIIEY. 1976. Sexual size
dimorphism in hawks and owls of North Amer-
ica. Ornithol. Monogr. No. 20.
Sor,L, R. R., & F. J. ROHF. 1981. Biometry, 2nd ed.
New York, W. H. Freeman and Co.
$TAHLECKER, D. W., & H. J. GRIESE. 1977ß Evidence
of double brooding by American Kestrels in the
Colorado high plains. Wilson Bull. 89: 618-619.
STEENHOF, K. 1983. Prey weights for computing
percent biomass in raptor diets. Raptor Res. 17:
15-27.
TOLAND, B.R. 1985. Double brooding by American
Kestrels in central Missouri. Condor 87: 434-436.
TORDOFF, H. B., & W. R. DAWSON. 1965. The influ-
ence of daylength on reproductive timing in the
Red Crossbill. Condor 67: 416-422.
WHALEY, W.H. 1979. The ecology and status of the
Harris' Hawk (Parabuteo unicinctus) in Arizona.
M.S. thesis, Tucson, Univ. Arizona.
1986. Population ecology of the Harris'
Hawk in Arizona. Raptor Res. 20: 1-15.