757 C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 22001133 SOUTSoHuEthAeSaTsEteRrnN NNaAtTurUaRliAstLIST 1V2o(4l.) :1725,7 N–7o6. 84 Time-activity Budgets of Waterfowl Wintering on Livestock Ponds in Northeast Texas Corey D. Mason1, R. Montague Whiting, Jr.2, and Warren C. Conway3,* Abstract - We developed individual time-activity budgets for Anas platyrhynchos (Mallard; n = 281), A. strepera (Gadwall; n = 251), and Aythya collaris (Ring-necked Duck; n = 144) wintering on livestock ponds in the Blackland Prairies Ecological Region of Texas in January and February, 2000 and 2001. Feeding (32–38%), locomoting (24–49%), and resting (10–36%) dominated the activity budgets for each species. Behaviors varied between years, probably due to the 3-fold increase in precipitation that raised water levels in livestock ponds. In 2000 and 2001, Mallards fed nearly 50% and 20% of their time, respectively, with comfort and resting occupying 60% in 2001. Gadwalls locomoted nearly 50% of their time each year, but increased surface feeding 2-fold in 2001. Finally, Ring-necked Ducks spent about a third of their time locomoting, another third resting, and the remainder subsurface feeding in 2001. Focal species activity budgets were generally similar to those developed throughout their ranges. Livestock ponds in northeast Texas provide small but regionally widespread habitats for wintering waterfowl. Future work should focus upon diet and landscape occupancy rates of waterfowl using these habitats during winter. Introduction Annual mid-winter aerial surveys performed by the Texas Parks and Wildlife Department (TPWD) have indicated relatively large numbers of ducks wintering on livestock ponds (i.e., stock tanks) in the Oak Woods and Blackland Prairies Ecological Regions (Gould 1975) of northeastern Texas. From 1997 to 2011, occupancy rates (i.e., ponds occupied by ≥1 duck) of small ponds (<0.81 h a) ranged from 11– 26%, and the estimated number of ponds increased from 277,890 to 299,583, while occupancy rates of medium-sized ponds (0.81–16.19 ha) ranged from 32–51%. Due to ongoing drought during the latter portion of this reporting period, the number of medium-sized ponds decreased from 34,835 in 1997 to 23,789 in 2011. During the same period (1997 to 2011), estimated numbers of ducks using these ponds ranged from 458,167 (2009) to 1,159,874 (2010), and the 15-year average was 759,069. The 3 most abundant species were Anas platyrhynchos L. (Mallard), A. strepera L. (Gadwall), and Aythya collaris Donovan (Ring-necked Duck) (mid-winter waterfowl surveys, 1997–2011; TPWD, Austin, TX, unpubl. data). In the northern United States, natural prairie wetlands are important to breeding waterfowl (Ball et al. 1995, Rumble and Flake 1983, Svingen and Anderson 1998). In similar habitats in the southern United States, natural wet-dry cycles 1Arthur Temple College of Forestry and Agriculture, Stephen F. Austin State University, Nacogdoches, TX. 75962. Current address - Texas Parks and Wildlife Department, 11942 FM 848, Tyler, TX. 75707. 2Arthur Temple College of Forestry and Agriculture, PO Box 6109, Stephen F. Austin State University, Nacogdoches, TX 75962. *Corresponding author - wconway@sfasu.edu. C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 758 drive plant and food production, thus providing key elements used by waterfowl during winter. Several studies have documented that wintering waterfowl typically use these habitats for feeding and loafing (Lee 1985, Quinlan and Baldassarre 1984). Particularly in Texas, livestock ponds may provide regionally important habitats during periods of extended drought. Waterfowl also use these ponds to escape hunting and other pressures commonly encountered on larger public reservoirs (see Crook et al. 2009). In northeast Texas, most livestock ponds are privately owned and experience little or no hunting pressure. Although small in size, these habitats may provide important habitat for waterfowl wintering in the region. In contrast to natural precipitation-filled prairie wetlands, livestock ponds have not received attention as potentially useful waterfowl habitats during winter. To address this information gap, we quantified diurnal time-activity budgets of Mallards, Gadwalls, and Ring-necked Ducks wintering on livestock ponds in the Blackland Prairies Region of Texas. Field Site Description This research was conducted in the Blackland Prairies Ecological Region of Texas (Gould 1975) on privately owned livestock ponds in Fannin and Hopkins counties in the very northern portion of the Blackland Prairies Region; Fannin County abuts the Red River and Oklahoma. Mean annual precipitation in the northern portion of the region is 100 cm, with the majority occurring during spring and summer, although amounts and distribution vary among months and years. Winters are generally mild with short periods of cold, wet weather during and following passage of cold fronts. The average frost-free season is March 15–November 17 (NRCS 2008). Due to climate, soils, and topography (i.e., flat to gently rolling hills; Gould 1975), land use is dominated by rowcrop agriculture and both improved and native pastures (NRCS 2008): <1% of the original vegetation and ephemeral wetlands remain regionally (TPWD 2005). In areas around our study ponds, the primary land use was cattle grazing. Methods We located potential study-site livestock ponds using ground surveys in December 1999 and 2000. We defined potential study-site ponds as those 1–20 ha in size, containing both surface water and the presence of ducks. Our choice of specific livestock ponds as study-sites was ultimately dependent upon landowner permission for access. We obtained access to 46 potentially suitable privately owned livestock ponds. All study-site ponds were <3 m deep, and typically contained emergent plants along the margins including Rumex spp. (docks), Polygonum spp. (knotweeds), and Eleocharis spp. (spike-rushes), as well as the floating-leaved aquatic species Nelumbo lutea Willd. (American Lotus), Ceratophyllum demersum L. (Coontail) and Lemna minor L. (Common Duckweed). We erected observation blinds prior to behavior sampling in early January 2000 and 759 C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 2001; blinds were positioned to minimize disturbing birds as observers arrived prior to sampling. We quantified time-activity budgets for Mallards, Gadwalls, and Ring-necked Ducks on 25 weekend days from 6 January–11 March 2000, and 10 weekend days from 9 January–25 February 2001. We used focal-individual sampling (Bergan et al. 1989, Crook et al. 2009, Turnbull and Baldassarre 1987) to collect behavior data during 3 diurnal periods (morning: sunrise–1030 h; midday:1031–1400 h; and afternoon: 1431 h–sunset; Bergan et al. 1989, Dwyer 1975, Lee 1985). We recorded the following behaviors at 15-second intervals for 5 minutes: (1) locomotion (i.e., swimming, walking, or flying); (2) resting (i.e., sleeping or loafing) on land; (3) resting on water; (4) comfort (i.e., body maintenance, including preening and bathing); (5) courtship (i.e., pair bonding displays, copulation, head-pumping); (6) agonistic (i.e., bill threats, chasing, and any other aggressive behaviors); (7) alert (i.e., cessation of all activities, assumption of upright posture); (8) surface feeding (i.e., feeding without tipping); (9) subsurface feeding (i.e., feeding by tipping or diving); and (10) land feeding (i.e., feeding out of water) (Crook et al. 2009, Jorde et al. 1984, Paulus 1984, Quinlan and Baldassarre 1984, Rave and Cordes 1993, Turnbull and Baldassarre 1987). A total of 5 observers collected behavior data during this study. Prior to data collection, observers trained together by watching domesticated Mallards on the Stephen F. Austin State University campus, until all observers were consistently able to correctly identify and classify each behavior. We randomly selected ponds for data collection on a given sampling day. Observers arrived at selected blinds 30 min before sunrise; if no focal species were present on a pond 1 h after sunrise, the observer moved to the nearest study site pond containing focal species. If >2 ducks were present, the observers began sampling the individual nearest the center of the group (Bergan et al. 1989, Rave and Baldassarre 1989, Turnbull and Baldassarre 1987). If only a single pair of ducks was present, the observers alternated recording behaviors for each individual. Likewise, if only a single individual was present, the observer continued to record that duck’s behavior. If a focal individual was lost from view or left the pond prior to the termination of the 5-min observation period, we terminated that sample, selected a new individual, and initiated a new observation period. After each 5-min observation period, the observer rested for 5 min, selected another individual, and began a new observation period. Each observer recorded all of his observations on standardized data sheets and entered the data into spreadsheets. During data collection, we instructed observers to record behaviors throughout the entire day. Consequently, there were instances in which observers repeatedly recorded behaviors on specific individuals. To prevent inflating sample sizes by repeatedly using data from the same individual birds and to maintain temporal independence among focal samples (the experimental unit in this study), we randomly selected a single focal sample for any individual during a given diurnal period (i.e., morning, midday, afternoon); thus, in our analysis we included no more than 3 focal samples on a specific duck each day. All other data were censured. C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 760 Data analysis We quantified individual time-activity budgets by calculating the proportion (%) of each behavior recorded during each 5-min focal sample (i.e., proportion of 20 instantaneous behaviors recorded at 15-sec intervals). We used a factorial multivariate analysis of variance (MANOVA) to examine differences in behaviors among species, between sexes, among diurnal periods, and between years (i.e., 2000 and 2001). We used MANOVA (Crook et al. 20009, Speights and Conway 2009) because individual behaviors within a focal sample are not independent. If differences occurred (P < 0.05) in MANOVA, we used univariate analysis of variance (ANOVA), followed by least squares mean separation if differences (P < 0.05) occurred during ANOVAs (Crook et al. 2009, Speights and Conway 2009). Results Due to the consistent presence of focal species, we collected behavior data on 8 of the 46 potentially available study site ponds; all were classified as small or medium-sized. A total of 676 focal samples were collected for Mallards (n = 281), Gadwalls (n = 251), and Ring-necked Ducks (n = 144), from 51 hours of focal observations. Ring-necked Duck behavior data were only collected in 2001. Behaviors varied among species (Wilks’ λ = 0.54; df = 18, 1196; P < 0.001), though all 3 species spent considerable time feeding (32–38%), locomoting (24–47%), and resting (10–36%) (Table 1). For both years combined, Mallards and Gadwalls spent approximately a third of their time feeding, as did Ring-necked Ducks in 2001 (Table 1). As expected, Ring-necked Ducks spent more time subsurface feeding (31%) than either Mallards (8%) or Gadwalls (9%), both of which spent 26–29% of their time surface feeding. In sum, all 3 species spent 32–37% of their time in feeding behaviors (Table 1). Mallards and Ring-necked Ducks generally fed in localized areas for extended periods of time, whereas Gadwalls normally fed while Table 1. Means (%), Standard Errors (SE), and F and P values resulting from univariate analysis of variance of wintering Mallard, Gadwall, and Ring-necked Duck behaviors measured on livestock ponds in northeast Texas, 6 January–11 March 2000 and 9 January–25 February 2001. Means followed by the same letter within the same row are not different (P > 0.05). Mallard Gadwall Ring-necked Duck (n = 218) (n = 251) (n = 144) Behavior Mean SE Mean SE Mean SE F P Locomotion 27 B 1.9 47 A 1.4 24 B 2.0 63.3 <0.001 Resting on land 5 A 1.4 2 B 0.7 <1 B 0.4 3.9 0.019 Resting on water 19 B 2.1 8 C 0.9 35 A 2.5 63.5 <0.001 Comfort 8 A 1.3 8 A 1.1 8 A 1.3 0.1 0.981 Courtship 1 A 0.3 <1 B <0.1 <1 B 0.1 6.1 0.002 Agonistic <1 A <0.1 <1 A <0.1 <1 A 0.2 0.7 0.507 Alert <1 A 0.2 <1 B <0.1 <1 B 0.2 12.6 <0.001 Surface feeding 29 A 2.4 26 A 1.5 1 B 0.4 44.9 <0.001 Subsurface feeding 8 B 1.3 9 B 1.0 31 A 2.1 47.1 <0.001 Feeding on land 1 A 0.5 <1 B 0.1 0 B 0.0 5.8 <0.001 761 C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 locomoting. Resting and comfort behaviors were more frequently noted in Ringnecked Ducks, which spent nearly 45% of their cumulative time in these behaviors (Table 1). In contrast, Gadwalls spent only 17% of their cumulative time in these behaviors (Table 1). There was a species by year interaction (Wilks’ λ = 0.86; df = 9, 598; P < 0.001); all subsequent analyses were performed within each species. Mallard behaviors varied between years (Wilks’ λ = 0.73; df = 9, 203; P < 0.001), but were similar between sexes (Wilks’ λ = 0.97; df = 9, 203; P = 0.791) and among time periods (Wilks’ λ = 0.88; df = 18, 406; P = 0.069). Subsequent analyses for Mallards were performed within each year, because behaviors were similar (P > 0.05) between sexes, among time periods, and between paired and unpaired individuals for both 2000 and 2001. Variation in Mallard activity budgets was driven exclusively by differences between years; Mallards spent over twice as much time feeding in 2000 (48%) as in 2001 (19%); almost 60% their time was dedicated to comfort and resting behaviors in 2001 (Table 2). Similarly, Gadwall behaviors varied between years (Wilks’ λ = 0.70; df = 9, 236; P < 0.001), but not between sexes (Wilks’ λ = 0.96; df = 9, 236; P = 0.276) or among time periods (Wilks’ λ = 0.90; df = 18, 472; P = 0.157). Subsequent analyses were performed within each year for Gadwall because behaviors were similar (P > 0.05) between sexes, among time periods, and between paired and unpaired individuals for both 2000 and 2001. Similar to Mallards, variation in Gadwall activity budgets was driven exclusively by differences between years; Gadwalls spent similar time locomoting (42–49%) each year, but spent more than twice as much time surface feeding in 2001 (Table 3). Analyses for Ring-necked Ducks were performed within 2001 only. As with the other species, Ring-necked Duck behaviors were similar among time periods (Wilks’ λ = 0.84; df = 16, 262; P = 0.107). Although overall behaviors varied (Wilks’ λ = 0.88; df = 8, 131; P = 0.041) between sexes and between paired and unpaired Ring-necked Ducks (Wilks’ λ = 0.88; df = 16, 232; P = 0.013), no individual behavior varied (P > 0.05) in subsequent ANOVAs (Table 4). Table 2. Means (%), Standard Errors (SE), and Type III F and P values resulting from univariate analysis of variance of wintering Mallard behaviors measured on livestock ponds in northeast Texas, 6 January–11 March 2000 and 9 January–25 February 2001. Means followed by the same letter within the same row are not different (P > 0.05). 2000 (n = 148) 2001 (n = 70) Behavior Mean SE Mean SE F P Locomotion 30A 2.3 21B 2.9 4.0 0.047 Resting on land 4A 1.5 8A 2.9 1.9 0.166 Resting on water 11B 1.9 37A 4.6 44.8 <0.001 Comfort 5B 1.3 14A 2.9 8.1 0.005 Courtship 1A 0.4 <1A 0.5 0.4 0.538 Agonistic <1A 0.1 <1A 0.0 2.2 0.139 Alert 1A 0.3 1A 0.3 0.2 0.627 Surface feeding 35A 3.0 17B 3.4 14.2 <0.001 Subsurface feeding 11A 1.9 1B 0.6 13.0 <0.001 Feeding on land 2A 0.7 <1A 0.1 3.1 0.081 C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 762 Discussion As with most other wintering waterfowl activity-budget studies (Clark and Whiting 1994, Crook et al. 2009, Paulus 1984), behaviors of all 3 focal species were dominated by feeding, locomoting, and resting/comfort activities (Table 5). Combined, these activities comprised >95% of all behaviors, and in sum, the activity budgets we documented do not deviate dramatically from other published wintering waterfowl activity-budgets, although time spent locomoting by Gadwalls was greater than others have reported (Table 5). Paulus (1988) estimated that nonbreeding waterfowl should allocate most of their time to feeding and resting (70–80%), with locomoting, and non-resting comfort behaviors, dominating the remainder of the time. However, locomoting alone exceeded 20% for each species’ activity budget in this study. Clearly, our focal species invest more time locomoting and Table 4. Means (%), Standard Errors (SE), and Type III F and P values resulting from univariate analysis of variance of wintering Ring-necked Duck behaviors measured on livestock ponds in northeast Texas, 9 January–25 February 2001. Means followed by the same letter within the same row are not different (P > 0.05). Males (n = 116) Females (n = 28) Behavior Mean SE Mean SE F P Locomotion 25 A 2.3 23 A 2.3 0.1 0.833 Resting on land 2 A 1.9 0 A 0.0 3.9 0.051 Resting on water 28 A 6.0 37 A 2.8 1.9 0.175 Comfort 13 A 4.4 7 A 1.3 2.6 0.107 Courtship <1 A 0.4 <1 A 0.1 2.1 0.148 Agonistic 0 A 0.0 <1 A 0.3 0.1 0.734 Alert 2 A 0.6 <1 A 0.2 2.2 0.140 Surface feeding 2 A 1.4 1 A 0.3 0.9 0.333 Subsurface feeding 29 A 5.4 31 A 2.3 0.1 0.760 Feeding on land 0 0.0 0 0.0 na na Table 3. Means (%), Standard Errors (SE), and Type III F and P values resulting from univariate analysis of variance of wintering Gadwall behaviors measured on livestock ponds in northeast Texas, 6 January–11 March 2000 and 9 January–25 February 2001. Means followed by the same letter within the same row are not different (P > 0.05). 2000 (n = 184) 2001 (n = 67) Behavior Mean SE Mean SE F P Locomotion 49A 1.7 42A 2.6 3.3 0.071 Resting on land <1B 0.5 5A 2.2 6.9 0.009 Resting on water 8A 1.1 6A 1.3 2.4 0.120 Comfort 10A 1.4 3B 1.0 10.0 0.002 Courtship <1A 0.1 0A 0.0 1.1 0.304 Agonistic <1A 0.1 <1A 0.1 3.7 0.054 Alert <1A 0.1 <1A 0.1 0.1 0.716 Surface feeding 19B 1.5 42A 3.0 51.7 <0.001 Subsurface feeding 12A 1.2 1B 0.6 22.1 <0.001 Feeding on land 0A 0.0 <1A 0.5 1.5 0.221 763 C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 comparatively less time feeding and/or resting than waterfowl in some other areas and habitats (Table 5). Reasons for these high rates of locomoting are unclear. Vehicular noise and gunshots were absent at our sites, and livestock did not use the ponds. As such, anthropogenic disturbances from boating, fishing, or hunting, all of which are often cited for increasing locomotion behaviors for wintering waterfowl within the region (Crook et al. 2009), did not influence behavio rs in this study. Ring-necked Ducks appeared to be relatively undisturbed, yet their behaviors, including locomoting, were almost identical to those recorded by Crook et al. (2009) on large (6800–75,000 ha) eastern Texas reservoirs (Table 5), where boat traffic apparently affected behaviors. In other regional studies, Mallards in a 728- ha bottomland hardwood forest spent almost 50% more time locomoting (Clark and Whiting 1994) than did birds in our study, which invested almost twice as much time in that activity than did Mallards in the Texas Panhandle (Lee 1985). Locomoting values in Lee’s (1985) study were very similar to those from Nebraska (Jorde et al. 1984) and Alabama (Turnbull and Baldassarre 1987) (Table 5). Thus, mechanisms driving elevated rates of locomotion seem to be inconsistent among studies, and may reflect responses to other, poorly defined cues. For Gadwalls, locomotion was the primary activity (47%) in this study, whereas it comprised only minor proportions of Gadwall activity-budgets in Louisiana (13%) and North Carolina (5%) (Table 5). In fact, our Gadwall locomotion estimate is greater than that found for any wintering waterfowl study (Paulus 1988:137–138). Consequently, Gadwalls in northeastern Texas spent much less time feeding (36%) Table 5. Summary of diurnal time-activity budgets of wintering Mallards, Gadwalls, and Ring-necked Ducks in the United States. BehaviorA Location Feeding Resting Locomoting Other Reference Mallard Nebraska 35 28 13 18 Jorde et al. (1984) Alabama 21 44 13 14 Turnbull and Baldassarre (1987) East Texas 20 22 43 13 Clark and Whiting (1994) Texas Panhandle 26 39 13 12 Lee (1985) Northeastern Texas 38 24 27 8 This study Gadwall North Carolina 75 17 5 3 Hepp (1972) North Dakota 70 8 8 9 Dwyer (1975) Louisiana 64 11 11 14 Paulus (1984) Northeastern Texas 35 10 47 8 This study Ring-necked Duck FloridaB 35 24 17 15 Hohman (1984) Mississippi 36 34 16 13 Christopher and Hill (1988) South CarolinaB 44 20 18 7 Bergan et al. (1989) East Texas 30 37 24 7 Crook et al. (2009) Northeastern Texas 32 35 24 8 This study APercentage of time performing individual behavior. BBehaviors approximated from figures. C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 764 than those in Louisiana (64%) and North Carolina (75%). Although Paulus (1988) noted that, for wintering waterfowl, time spent feeding and resting are usually inversely related, and our results support this notion for Mallards and Ring-necked Ducks, this was not the case for our observations of Gadwalls. In fact, our Gadwall data suggest that feeding was inversely related to locomoting rather than resting, and our estimated time spent resting (10%) is at the bottom of the range (10–50%) for nonbreeding Anatids (Paulus 1988). Our data collection protocol may have underestimated feeding time and over-estimated locomoting time by Gadwalls, as they frequently fed by dabbling or filtering as they swam, or by occasionally tipping for a few seconds and then moving along. Gadwalls rarely performed extended feeding bouts, unlike in many other dabblingduck studies, that described individuals specifically gathering or consuming foods (Paulus 1982). However, our locomotion estimates were similar between years, even when specific foraging behaviors (i.e., surface vs. subsurface) changed dramatically between years (see Table 3), indicating that locomoting was in fact a dominant behavior for Gadwalls using these livestock ponds. Interestingly, Gadwall feeding behavior was more similar to that of Aix sponsa L. (Wood Duck) (Clark and Whiting 1994, Drobney and Fredrickson 1979) than typical Gadwall feeding behavior described by other researchers (Dwyer 1975, Paulus 1982). Such behavioral similarities to Wood Ducks have not been previously reported. Gadwalls were relatively widely distributed and commonly observed on our study-site ponds, and it is possible that they were feeding elsewhere, primarily at night. No ducks were ever present on study ponds when observers arrived before daylight, and all always left by dark. Nocturnal feeding behavior is often cited as a response to hunting pressure or other disturbances (see Korschgen and Dahlgren 1992). Therefore, all 3 focal species, but especially Gadwalls, may have been pursuing nocturnal foraging opportunities elsewhere, and used the livestock ponds diurnally for supplemental foraging. Future work should examine food availability and quality to more clearly examine the relationship between foraging behaviors and nutritional food values to more clearly characterize the importance of regional livestock ponds for wintering waterfowl. Similar to other studies of Mallards (Turnbull and Baldassarre 1987), Gadwalls (Paulus 1984), and Ring-necked Ducks (Crook et al. 2009, Hohman 1984), behaviors were similar between sexes for all 3 species. Perhaps activity budgets were similar because 1) both sexes have similar nutrient requirements (Hohman 1984, Paulus 1984), and 2) livestock-pond quality was not impacting activity budgets between sexes or between paired and unpaired individuals. Although some studies have noted between-sex differences (Bergan et al. 1989, Clark and Whiting 1994, Jorde et al. 1984), such differences may be a result of intersexual aggression and male dominance (Hohman 1984) or that females, which are smaller than males, are at a thermal disadvantage (Bergan et al. 1989). However, agonistic and courtship behaviors were infrequently observed during this study (<1% for all species). Variation in activity budgets among diurnal periods is inconsistent among studies; some researchers have found variation (Clark and Whiting 1994, Hohman 1984, Paulus 1984), and others have not (Turnbull and Baldassarre 1987). There is 765 C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 no strong, consistent pattern of daily activities among species or studies. Generally, Mallards rested in the morning and fed in the afternoon (Clark and Whiting 1994, Jorde et al. 1984), whereas Gadwalls fed early and late and rested at midday (Paulus 1984). Ring-necked Ducks in Florida were most active (feeding, locomoting, alert) in morning (Hohman 1984), but those in South Carolina fed throughout the day, and rested and locomoted most in the afternoon (Bergan et al. 1989). No such diurnal variation was observed in this study, where activity budgets were consistent among all time periods. Yearly variation in activity budgets is not uncommon, because changing habitat conditions and weather patterns can dramatically influence behavior from year to year. In this study, differences between years was the primary factor influencing variation in activity budgets for Mallards and Gadwalls; we contend that this factor was strongly related to water availability during the study. In 1999, the study area received only 71 cm of rainfall, compared to 139 cm in 2000; temporal distribution of the precipitation was also a factor in altering water levels in the livestock ponds during our study. There were 30.0 cm of precipitation during November–December 1999 and January–February 2000, whereas 83 cm of precipitation were experienced during the same months in 2000–2001—nearly a 3-fold difference in precipitation (Weather Station Records, US Army Corps of Engineers, Sulphur Springs, TX, unpubl. data). Precipitation in late fall and early winter of 2000–2001 flooded pastures and agricultural fields, and allowed ducks to disperse to these areas as well as to recently flooded wetlands adjacent to nearby reservoirs. Although numbers of focal samples collected in 2001 were <50% of those recorded in 2000, these decreases do not necessarily reflect a decrease in the number of waterfowl wintering regionally. In fact, there were more Mallards and Gadwalls observed during mid-winter (January) waterfowl surveys in 2001 than 2000 in the Oak Woods and Blackland Prairie Ecoregions —2000: Mallards 284,907; Gadwalls 203,484; 2001: Mallards 453,762, Gadwalls 245,126 (mid-winter waterfowl survey 2000, 2001; TPWD, Austin, TX, unpubl. data). We presume that Gadwalls and Mallards were using other wetland types more frequently than livestock ponds in 2001. We assume that the birds moved to other potentially suitable habitats, and that such movements may have accounted for the changes in Mallard and Gadwall behaviors in 2001. In 2000, Mallards spent nearly 50% of their time in feeding behaviors, but in 2001, they spent <20% of their time feeding and nearly 50% of their time resting (Table 2). Conversely, Gadwalls more than doubled the amount of time they spent surface feeding from 2000 to 2001, and increased locomotion from 42 to 49% (Table 3). These differences in Mallard and Gadwall feeding behaviors may be related to the 3-fold increase in precipitation from 2000 to 2001. The drought of 1999 and early 2000 reduced water levels in the ponds and nearby reservoirs, and in 2000, the only water available for ducks in Hopkins and Fannin counties was in ponds and reservoirs; thus, potential feeding sites were limited, and the birds were concentrated. In 2001, Mallards were widely distributed, and as we traveled between study sites searching for focal species, we commonly saw Mallards feeding C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 766 in flooded fields and pastures and large flocks of Gadwalls and Ring-necked Ducks on nearby reservoirs. Mallards typically prefer shallow-water feeding-sites such as the flooded fields and pastures noted in 2001. For example, Mallards wintering in Alabama foraged in shallow impoundments and often made evening feeding flights to flooded agricultural fields (Turnbull and Baldassarre 1987). Similar to this study, Mallards spent time in non-feeding behaviors on a deep impoundment and a river (Turnbull and Baldassarre 1987). In the Texas Panhandle, Mallards spent more time feeding in shallow-water terraced pits than in steep-sided pits and open-water lakes combined, although they spent approximately the same amount of time resting in each of the 3 habitat types (Lee 1985). In 2001, the increased surface feeding that we observed for Gadwalls may have been due to increased food availability brought by run-off into recently inundated areas around the ponds; we rarely observed subsurface feeding in 2001. Subsurface feeding requires more energy and, with increased food supply, may simply not have been necessary (sensu Goldstein 1988, Lovvorn and Gillingham 1996). Similarly, vegetation available for subsurface feeding in 2000 may not have been available in 2001 due to increased water depths and/or changes in plant species composition. Increased food availability in 2001 may have resulted in Gadwalls investing more time in feeding and resting on land, thus reducing time spent incomfort activities. Regional livestock ponds clearly provide wintering habitats for our focal species, as evidenced from mid-winter surveys, where >500,000 individuals of our focal species typically occur in the Oak Woodlands and Blackland Prairie Ecoregions during winter. However, Anas clypeata L. (Northern Shoveler) and A. crecca Gmelin (Green-winged Teal) also regularly used ponds, as did 9 other species (Mason 2002). Understanding how ducks use various habitats is especially important before making decisions about how to create and manage man-made waterfowl habitat, which are usually developed as shallow-water features. In contrast, livestock ponds are designed as deeper water bodies. However, our results suggest that small livestock ponds are also necessary for diurnal foraging and resting, and may be critical in dry years, when natural or man-made shallow-water habitats are limited or absent on the landscape. To more clearly characterize the regional value of these ponds, future work should focus upon waterfowl feeding habits, livestock-pond food production, and nutritive quality of those items as related 1) to water level, and 2) occupancy by wintering waterfowl. Such work should facilitate development of more specific conservation and management guidelines for these regional wintering waterfowl habitats. Acknowledgments This study was funded by the Arthur Temple College of Forestry and Agriculture, Stephen F. Austin State University. We thank S. Cordts, C. Frentress, W. Johnson, K. Kraai, D. Morrison, and B. Sullivan for their ideas and assistance. R. Mangham, J. Laing, S. Locke, and S. Crook assisted with gathering behavioral data. S. Leeper assisted with manuscript preparation. 767 C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 Literature Cited Ball, I., R. Eng, and S. Ball. 1995. Population density and productivity of ducks on large grassland tracts in north-central Montana. Wildlife Society Bulletin 23:767–773. Bergan, J.F., L.M. Smith, and J.J. Mayer. 1989. Time-activity budgets of diving ducks wintering in South Carolina. Journal of Wildlife Management 53:769–776. Christopher, M.W., and E.P. Hill. 1988. Diurnal activity-budgets of nonbreeding waterfowl and coots using catfish ponds in Mississippi. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 42:520–527. Clark, H.B., and R.M. Whiting, Jr. 1994. Time budgets of Mallards and Wood Ducks wintering in a flooded bottomland hardwood forest. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 48:22–30. Crook, S.L., W.C. Conway, C.D. Mason, and K.J. Kraai. 2009. Winter time-activity budgets of diving ducks on eastern Texas reservoirs. Waterbirds 32:548–558. Drobney, R.D., and L.H. Fredrickson. 1979. Food selection of Wood Ducks in relation to breeding status. Journal of Wildlife Management 43:109–120. Dwyer, T.J. 1975. Time budgets of breeding Gadwalls. Wilson Bulletin 87:335–343. Goldstein, D.L. 1988. Estimates of daily energy expenditures in birds: The time-energy budget as an integrator of laboratory and field studies. American Zoologist 28:829–844. Gould, F.W. 1975. The Grasses of Texas. Texas Agricultural Experiment Station, College Station, TX. 112 pp. Hepp, G.R. 1982. Behavioral ecology of waterfowl (Anatini) wintering in coastal North Carolina. Ph.D. Dissertation. North Carolina State University, Raleigh, NC. 155 pp. Hohman, W.L. 1984. Diurnal time-activity budgets of Ring-necked Ducks wintering in Central Florida. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 38:158–164. Jorde, D.G., G.L. Krapu, R. Crawford, and M.A. Hay. 1984. Effects of weather on habitat selection and behavior of Mallards wintering in Nebraska. Condor 86:258–265. Korschgen, C.E., and R.B. Dahlgren. 1992. Human disturbances of waterfowl: Causes, effects, and management. Fish and Wildlife Leaflet 13.2.15. Lee, S.D. 1985. A time-budget study of Mallards on the Texas High Plains. M.Sc. Thesis. Texas Tech University, Lubbock, TX. 32 pp. Lovvorn, J.R., and M.P. Gillingham. 1996. Food dispersion and foraging energetics: A mechanistic synthesis for field studies of avian benthivores. Ec ology 77:435–451. Mason, C.D. 2002. Diurnal time budgets of Mallards and Gadwalls wintering on livestock ponds in Northeast Texas. M.Sc. Thesis. Stephen F. Austin State University, Nacogdoches, TX. 62 pages. Natural Resource Conservation Service (NRCS). 2008. Ecological Site Description–Texas Blackland Prairie, Northern Part. Available online at http://esis.sc.egov.usda.gov/ ESDReport/ fsreport.aspx?id=RO86AY196TX&rptLevel=all&a. Accessed 6 June 2011. Paulus, S.L. 1982. Feeding ecology of Gadwalls in Louisiana in winter. Journal of Wildlife Management 46:71–79. Paulus, S.L. 1984. Activity of nonbreeding Gadwalls in Louisiana. Journal of Wildlife Management 48:371–380. Paulus, S.L. 1988. Time-activity budgets of nonbreeding Anatidae: A review. Pp. 135–152, In M.W. Weller (Ed.). Waterfowl in Winter. University of Minnesota Press, Minneapolis, MN. 624 pp. Quinlan, E.E., and G.A. Baldassarre. 1984. Activity budgets of nonbreeding Green-winged Teal on playa lakes in Texas. Journal of Wildlife Management 48:838–845. C.D. Mason, R.M. Whiting, Jr., and W.C. Conway 2013 Southeastern Naturalist Vol. 12, No. 4 768 Rave, D.P., and G.A. Baldassarre. 1989. Activity budget of Green-winged Teal wintering in coastal wetlands of Louisiana. Journal of Wildlife Management 53:753–759. Rave, D.P., and C.L. Cordes. 1993. Time-activity budget of Northern Pintails using nonhunted rice fields in Southwest Louisiana. Journal of Field Orni thology 64:211–218. Rumble, M., and L. Flake. 1983. Management considerations to enhance use of stock ponds by waterfowl broods. Journal of Range Management 36:691–694. Speights, J.R., and W.C. Conway. 2009. Wintering Yellow-bellied Sapsucker time-activity budgets in East Texas bottomland hardwood forests. The Wilson Journal of Ornithology 121:593–599. Svingen, D., and S.H. Anderson. 1998. Waterfowl management on grass-sage stock ponds. Wetlands 18:84–89. Texas Parks and Wildlife Department (TPWD). 2005. Texas Comprehensive Wildlife Conservation Strategy 2005–2010. Austin,TX. Turnbull, R.E., and G.A. Baldassarre. 1987. Activity budgets of Mallards and American Wigeon wintering in East-central Alabama. Wilson Bulletin 99:457–464.