White-tailed deer (Odocoileus virginianus; Artiodactyla: Cervidae) is distributed throughout the Americas, from Canadian forests, to coniferous and xerophytic forests in the US, in most forests in Mexico and even in portions of South America¹. It is widely hunted in Mexico², and is raised in Wildlife Conservation Management Units (Unidades de Manejo para la Conservación de la Vida Silvestre - UMA) to produce trophies, meat, skin, brood stock, and ornaments, among other products³.

In the wild, O. virginianus is an opportunistic selective herbivore, foraging a selection of plant parts (e.g. shoots, fruits, leaves, bark, and seeds), especially those with high nutritional value⁴. When the dry season occurs in deciduous tropical forests plant abundance decreases and their nutritional quality diminishes⁵. Under these circumstances, O. virginianus can experience deficiencies in development, such as a lower than standard weight, become prone to disease and limit its reproduction⁴. These same responses are often observed in captive O. virginianus. Captive deer, fed diets based on sheep and commercial deer feeds as well as alfalfa⁶, produce single rather than twin births, have low birth weight offspring, and longer intervals between births⁷.

Adult deer require 5.5 to 9 % crude dietary protein for adequate physiological development^8,9. Protein requirements may be related to ontogeny⁹, since captive fawns require between 13 and 20 % protein for adequate development, while, for optimal antler development, 15 to 18 % protein is required⁹. Females require from 11 to 18 % protein in pre-breeding, mating, pregnancy, lactation, and to increase offspring count¹⁰. Diet diversification in O. virginianus UMAs is imperative to complement basic feed nutritional value and improve productive characteristics¹¹. If animal feed preferences, nutrients contained in preferred plants and the nutritional requirements of animals at given weights can be interrelated, then animal productive behavior can be estimated¹¹.

Estimates of the nutritional content of plants consumed by wild deer have been done using various methodologies^12,13, but none have been done for captive deer. Cafeteria tests allow quantification and analysis of how animals modify dietary behavior to balance their nutritional needs. Essentially a multiple choice test, animals are offered one or several plants and their nutritional preferences documented¹⁴. The present study objective was to use a cafeteria test to quantify the dietary preferences of captive O. virginianus offered eight plants as feed.

The study was carried out at the El Pochote UMA (Secretaría de Medio Ambiente y Recursos Naturales registry: UMA-IN-CR-0122-VER/og), located in Ixtaczoquitlán municipality, in the state of Veracruz, Mexico (coordinates: 18°52’13.70” N; 97°02’59.97” W; 1,137 m asl). Regional climate is predominantly semi-warm humid (Cwa) with abundant summer rains, an average annual temperature of 18 to 24 °C, and average annual rainfall of 1,900 to 2,600 mm. Vegetation near the UMA consists of remnant semi-evergreen tropical forest and secondary vegetation.

Experimental animals were two-year-old deer (3 males and 3 females, n = 6), all healthy and with similar body conditions. The cafeteria feeding trial was done over a 60-d period, that is, three replicates of 5 d feeding with the eight selected plants, followed by a 15-d evaluation. Feeding with the selected plants was done for five consecutive days at 0900 h. Independent feeders were randomly distributed within the pen, and 1 kg fresh material (leaves, shoots and green branches) from each of the tested plants placed in separate feeders (Table 1). To reduce animal subjectivity (deer tend to repeat feeding behaviors), feeder positions were changed daily. After 2 h, the feeders and the remaining plant material were removed from the pen. Intake was quantified with the equation consumption = grams material offered - grams material rejected.

Proximate analyses were done of the eight tested plant species. Three samples of 100 g of mixed material were collected from each plant and incinerated for 2 h at 600 °C. Organic matter, ash, °Brix, pH and acidity were estimated; crude protein was quantified with the Kjeldahl method (N x 6.25) and ether extract in a Soxhtel extractor¹⁵. The intake and physicochemical analysis data were analyzed with descriptive statistics using a central tendency. Intake levels by animal were analyzed with an analysis of variance (ANOVA) and a Tukey means test (α=0.05). A partial least squares (PLS) regression analysis was applied in which the dependent variable was intake per plant species, the categorical variables were the eight plants, and the predictor variables were each plant’s physicochemical characteristics. All analyses were run with the Infostat ver. 2017 software.

The average intake results (Table 1) showed Bursera simaruba to have the highest coefficient of variation and the lowest average intake. The ANOVA identified Zapoteca aculeata, Bidens pilosa, Parthenium hysterophorus and Pennisetum purpureum as having the highest intake (correlation coefficient: R²= 0.96, coefficient of variation= 5.94; P<0.05; Table 2). These levels exceeded those of the other evaluated plants (Tukey: minimum significant difference = 135.68 g, error= 2204.01, gl= 16; Figure 1). This is supported by the coefficients of variation, since only these four plants were clearly preferred by the animals. The tested plant species varied in terms of protein, fiber and °Brix (Table 3). The PLS regression analysis explained 61.7 % of the correlation for intake preference of V. farnesiana, B. pilosa, Z. acuelata and S. officinarum, which was related to fiber and protein contents and °Brix level (Figure 2).

Figure 1 Tukey means test results identifying plants with highest intake by O. virginianus 

Figure 2 Relationship of plant physicochemical characteristics to intake by O. virginianus 

Of the tested plants, Z. aculeata, B. pilosa, P. purpureum, P. hysterophorus and S. officinarum were preferred by the captive O. virginianus. That these include two herbaceous plants and a grass is of note since wild deer consume mostly shrubs and trees, choose herbaceous species only seasonally, and consume few grasses throughout the year^16,17. Voluntary consumption of bushy, herbaceous and grassy plants reflects nutritional need^18,19, and is focused on species with the best physicochemical characteristics²⁰, such as carbohydrates (°Brix) and fiber, both vital to digestibility²¹.

All eight tested plant species meet deer protein requirements according to ontogenic stage. To reach above-average weight male deer require 15 % dietary protein²¹, and females require 13 %²². In young males, optimal growth requires from 13 to 16 % protein, while 20 % will augment their reproductive activity²².

These results suggest that at least five of the tested plant species could be used to diversify the diet of captive O. virginianus, which represents more dietary options for UMAs in this region. In addition, the tested plants have physicochemical characteristics that make them apt for use as deer feed while meeting the productive and reproductive requirements of O. virginianus.