28 
nesota contributed 5 to 8% each (Table A-65), The EMU 
contributed 4%. 
Age Differences in Arrival Patterns in Texas 
Earlier in the analysis, it was determined that a higher 
proportion of the direct recoveries of immature doves 
occurred outside the State of banding than that for adults 
(P < 0.01). Thus, it was deduced that immatures departed 
their natal area before adults. Another way to examine the 
movements of adult and immature doves is to determine 
whether immatures from a given area arrive at a point in 
their southward migration earlier than adults. 
North Texas was used as the recovery area. The number 
of direct band recoveries by 10-day intervals during Sep- 
tember and October from doves banded outside Texas was 
sufficient to test the hypothesis that adult and immature 
doves have similar chronological recovery patterns in North 
Texas. A series of log likelihood ratio tests of independence 
(Sokal and Rohlf 1969) was used to test the hypothesis 
(Table A-68). 
Direct band recoveries from most States were insufficient 
for testing on an individual State basis. Of the States tested, 
no differences were detected except for Kansas (Table A-68, 
Test No. 1). 
Direct recoveries of banded adults from all States out- 
side Texas were pooled and compared with the pooled data 
for immatures. The test of independence (Test No. 2) indi- 
cated that the chronological recovery pattern of adults was 
different (P < 0.01) from that of immatures. To determine 
which periods were accounting for most of the variation, 
we removed the 1-10 September period and tested the 
remaining five periods for independence (Test No, 3). No 
difference was detected. These five periods were pooled and 
tested against the first 10-day period (Test No. 4). A dif- 
ference (P < 0.01) was detected. The 1-10 September 
period accounted for about 77% of the total chi-square 
value or 77% of the aberration. 
The States were then combined into two groups. Mon- 
tana, North Dakota, Minnesota, Wisconsin, Illinois, lowa, 
South Dakota, Nebraska, and Wyoming were classified as 
northern States. New Mexico, Colorado, Oklahoma, 
Kansas, Missouri, Arkansas, and Louisiana were classified 
as southern States. Data were combined within each group 
and the hypothesis tested that immature and adult mourn- 
ing doves have similar chronological recovery patterns in 
North Texas. 
No difference was detected between recovery patterns 
for adult and immature doves from northern States (Test 
No. 5). No difference was detected between recovery 
patterns for adult and immature doves from southern States 
(Test No. 6), However, to be consistent with earlier test- 
ing, the first 10-day period was removed and the remain- 
ing five periods were tested for independence (Test No. 7). 
No difference was detected so the five periods were com- 
bined and tested against the 1-10 September period (Test 
No. 8). Results indicated a rejection of the null hypothesis 
(P < 0.01); therefore, the 1-10 September period 
accounted for most of the variation (about 82% ). This indi- 
cates that immatures arrived earlier than adults, were more 
vulnerable to the gun than adults early in the season, or 
a combination of both. 
If immature doves were considerably more vulnerable 
than adults during early September, and the vulnerability 
difference between the two decreased as the season pro- 
gressed, a recovery pattern similar to that observed could 
be expected. This potential explanation was tested indirectly 
by comparing the chronological recovery pattern in North 
Texas of adults banded in North Texas with immatures 
banded in North Texas (Test No. 9). These data failed to 
indicate a vulnerability bias; therefore, we conclude that 
the change in proportions was the result of early arriving 
immatures. 
Similar procedures were followed for recoveries in South 
Texas: however, no chronological differences were detected. 
Inconsistent season dates in South Texas were not conducive 
to such testing. 
Chronology of Recoveries in Mexico 
The chronology of direct recoveries in Mexico was 
examined for doves banded in the CMU during 1967-74. 
Chronological recovery patterns were used to depict the 
timing of hunting pressure and to approximate arrival dates. 
In Mexico, the dates of the mourning dove hunting sea- 
sons varied considerably during the span of this study 
(Table A-4). Seasons opened as early as 1 August, and as 
late as 1 November. Seasons encompassed October in 5 of 
8 years and November in 6 of 8 years. However, it is recog- 
nized that hunting in Mexico was not necessarily confined 
to the official season. This factor was the reason why we 
included all Mexican recoveries from August through May 
in this report. We believe that the chronological pattern 
of recoveries in Mexico was not strongly influenced by the 
hunting season dates prescribed by the law and, therefore, 
that the recovery dates represented the arrival pattern of 
CMU doves. However, it is recognized that if hunting sea- 
sons uniformly had opened earlier, a larger proportion of 
the recoveries might have occurred during September. 
Ten-day periods in which direct recoveries were reported 
from Mexico were tabulated for adult and immature doves 
(Table A-69), Of the 184 direct recoveries of adults and 511 
of immatures, 162 (88%) and 457 (89 % ) were identifiable 
to the 10-day period of recovery. 
The recovery patterns indicate that both adult and imma- 
ture doves from the CMU began arriving in southern har- 
vest areas of Mexico during late September and early 
October; there was a marked increase during 11-20 
October. Recoveries during this period were 3 times that 
of 1-10 October, and were the highest of all 10-day periods 
(21 and 67 recoveries for adults and immatures, respec- 
tively). Recoveries for adults remained high through 1-10 
November; they averaged about 16 recoveries per 10-day 
period, decreasing to about 6 recoveries per 10-day period 
