Panacea or frustration?

Infrared-triggered camera systems have added a whole new dimension to hunting, as well as research and management, of white-tailed deer. Compared to other observational techniques, camera systems are less disruptive and relatively inexpensive. Most importantly, cameras are easy to use, provide permanent documentation of “captured” animals, and permit gathering information at night and during nasty weather.

Maybe it’s my advancing age (Let’s call it experience.), or maybe my keen interest in trapping, whatever, I find capturing the image of a nice buck using a trail camera just about as satisfying as I do shooting one.

According to Auburn University researchers led by Clint McCoy, use of “remote photography” as a deer research and management tool has increased in popularity. Even private landowners are using camera systems to determine fawn-doe ratios, adult sex ratios, buck quality, and population size in order to make deer management decisions.

Probably the most frequent application of camera surveillance is by hunters, many who use remote cameras to search for that better-than-average buck. Certainly, the popular literature abounds with stories detailing hunter success credited to trail cameras. However, despite intensive camera surveillance, seldom do you read about the failure to camera-detect big bucks that just suddenly seemed to materialize out of nowhere. Or, worse yet, what about the monstrous pre-hunt-detected bucks that fails to show once the hunting season starts?

In either case, the unexpected appearance or untimely disappearance of certain camera-detected deer suggests that captured images during any given time period might not accurately represent the sex/age composition nor size of the local deer population — especially when using bait as an attractant.


Scientists have developed camera-survey techniques they believe reliably determine the composition and size of deer populations. However, these surveys commonly use bait to attract deer for camera detection.

In the early 1990s, researchers led by Mississippi State University’s Harry Jacobson used infrared-triggered cameras and monitors to census deer on a 15.6 square-mile area in Mississippi. Using bait to attract deer, and by determining the number of branch-antlered buck, spike bucks, does, and fawns, they reported that the camera technique could provide reliable estimates of deer population size — with a density of about 4 cameras per square-mile.

In other studies, researchers determined that camera survey estimates of deer were just as reliable as helicopter counts and observational road surveys of deer in south Texas and Florida, respectively.

Texas researchers Ben Koerth and James Kroll found corn the most attractive bait and suggested that the camera technique provided precise estimates of deer sex and age ratios in one week or less. However, later they determined that no single month provided best estimates of deer sex and age ratios when using bait as an attractant. Instead, they recommended a multi-stage survey in order to reliably estimate deer sex and age ratios.

Since all deer may not use bait equally, McCoy and his coworkers speculated that seasonal variations in the sex and age of camera-detected deer at bait piles may bias estimates and contribute to faulty herd management decisions. As a result, they designed and conducted a study to compare the sex, age, and abundance of deer photographed at baited sites versus those photographed along trails and at random sites. They also examined seasonal fluctuations in feeder use by deer to determine the best time to conduct a camera survey.

The Study

McCoy and fellow researchers conducted their study in a one square-mile enclosure in east-central Alabama. The high-density enclosed population (about 145 deer) was presumably maintained in good health with a high quality diet via food plots and supplemental feeding of a commercial ration. Due to selective harvesting, the adult sex ratio favored males 2:1.

The investigators monitored cameras placed at random, along heavily used game trails, and at feed stations from early September 2008 until early March 2009. Each sampling period consisted of 1 week, for a total of 19 sampling periods.

The number of fawns, does, and bucks was recorded in each photo. Based on antler and body characteristics, bucks were categorized as yearling (1.5 yr.), adult (2.5-3.5 yr.), and mature (older than 3.5 yr.). Four seasons were used to determine any seasonal effects: fall (11 Sept-31 Oct), pre rut (1 Nov-26 Dec), rut (6 Jan-7 Feb), and postrut (8 Feb-5 Mar).

Hence, the study was designed to simulate an actual camera survey that would be conducted to determine deer population sex and age composition. As such, the goal was to determine if the frequency of deer at feeders was similar to the true population structure, which was presumably provided by the random camera sites.

Seasonal Variation

In all, the researchers counted 5,322 deer in 3,972 distinct photos. Understandably, far more deer were photographed at feeders (75.4 percent) than at random (8.7 percent) or trail (16.0 percent) sites.

Also, more deer were observed at feeders during the fall period than during the other 3 seasons. By comparison, similar numbers of deer were photographed at random and trail sites each season.

Interestingly, doe use of feeders did not vary from that expected (i.e., random sites) during any season, but use by fawns, yearling bucks, and older bucks did. Since adult and mature bucks did not differ in their response, data for bucks was combined into one category.

Although fawns made limited use of feeders during fall, they were overrepresented in feeder photos during the rut, prerut, and postrut. In fact, fawns were photographed at feeders 5.5 times more often than expected during the rut.

Yearling bucks made greater use of feeders than expected during fall. However, during the prerut and rut periods, yearling and adult bucks were underrepresented in photographs taken at feeders.

In general, cameras placed at feeders did not generate the expected deer population structure, as determined at random sites, during any single time period. While camera surveys conducted at feeders during both the fall and postrut periods provided reasonable estimates of adult population structure, they did not provide reliable estimates for fawns.

Bryan Kinkel and Grant Woods reported somewhat similar seasonal variations resulting from camera surveys in an earlier Deer & Deer Hunting article. Initially, using trail-based cameras, they identified 9 deer on an unfenced 500-acre property in September. However, a minimum of 54 deer were detected from October through January — due largely to the attractiveness of an excellent acorn crop.

“This shift,” says Krinkel and Woods, “exemplifies an important point about deer home ranges: Too many hunters think a deer’s home range is a large circular or elliptical circuit it travels daily. But…as food conditions change, deer might shift to the opposite end of their home range, which, depending on habitat conditions, might be a half-mile or a mile or more away.”

Recruitment Rates

It is noteworthy that fawns in the Alabama study area are normally born in August, so they probably were not very mobile during the fall survey — accounting for few fawn photos taken at feeders during fall.

Conversely, fawns were overrepresented at feeders during the prerut, rut, and postrut. This means annual recruitment rates (fawns per doe) are likely to be biased low in fall but biased high thereafter — at least in southern areas where most fawns are probably born in July and August.

For example, in the Alabama study, estimates of recruitment using random sites were similar throughout the prerut, rut, and postrut periods, ranging from 0.72 to 0.80 fawns per doe. However, during the same time, estimated recruitment rates at feeders ranged from 1.13 to 2.84 fawns per doe.

Quite likely, such a survey in a more northerly area, where most fawns are born in June, would produce different results. That is, fawns probably would be photographed more frequently than expected at feeders during fall as well as throughout the prerut to postrut period.

In the North, wherever fawn per doe ratios are assessed at bait piles during the annual rut, as they often are here in northern Michigan, the data could yield erroneously high annual recruitment rates — in some cases 2 to 3 times higher than actually occurred. However, in recent years, deer hunting camp surveys conducted in Michigan’s Upper Peninsula have produced fawn per doe ratios of around 0.4 to 0.6, which seem to be realistic.

Therefore, regardless of the geographic region, it appears that camera surveys conducted at bait probably are not suitable locations for estimating annual recruitment rates.

Sex Ratios

Apparently, adult sex ratio estimates from camera surveys at feed stations (and presumably bait piles) may be inaccurate during any time period other than fall. Both adult does and bucks presumably feed heavily during the fall period in preparation for the breeding season.

Likewise, it certainly isn’t surprising that adult males were underrepresented in photos at feed stations during the rut. During the rut, adult males typically eat less, expend more energy in travel, and lose body weight.

However, I find it interesting that bucks were less frequently photographed than expected at all 3 survey sites (feeders, random, and trails) during the prerut period. The researchers speculate that fattened adult bucks exhibit a suppression in activity during the prerut period, possibly in anticipation of excessive energy demands associated with the breeding season. In fact, several other studies have demonstrated that movement rates of bucks tend to increase by nearly 30 percent from the prerut to rut periods.


Data from the Alabama study indicates that camera surveillance at bait sites generally does not provide reliable information concerning deer population structure nor annual recruitment rates. While random camera locations probably provide the best index to deer herd composition, such sites yield few deer photos, especially in areas of low deer density. Therefore, cameras positioned on game trails may provide the best data, because they produce population estimates similar to random sites and roughly twice as many deer photos. However, the number of cameras used, or the amount of time they are deployed, may need to be increased along trails in areas of low deer density.

This is not to say that camera surveys at bait will not produce favorable results at certain times in some areas. Results will undoubtedly vary regionally, depending upon timing of the breeding season and its effects on deer activity. Even annual variations in availability of important late summer and autumn food sources, within the same area, are likely to influence camera survey results at bait.

Meanwhile, deer-wise hunters will continue to find trail-cameras a panacea in their quest for the better-than-average buck — with or without use of bait. However, those hunters who fail to understand how deer behavior is driven by seasonal changes associated with breeding and food requirements are likely to achieve little more than trail camera frustration.