Density‐dependent behavior underpins white‐tailed deer ( Odocoileus virginianus ) theory and management application in North America, but strength or frequency of the phenomenon has varied across the geographic range of the species. The modifying effect of stochastic environments and poor‐quality habitats on density‐dependent behavior has been recognized for ungulate populations around the world, including white‐tailed deer populations in South Texas, USA. Despite the importance of understanding mechanisms influencing density dependence, researchers have concentrated on demographic and morphological implications of deer density. Researchers have not focused on linking vegetation dynamics, nutrition, and deer dynamics. We conducted a series of designed experiments during 2004–2012 to determine how strongly white‐tailed deer density, vegetation composition, and deer nutrition (natural and supplemented) are linked in a semi‐arid environment where the coefficient of variation of annual precipitation exceeds 30%. We replicated our study on 2 sites with thornshrub vegetation in Dimmit County, Texas. During late 2003, we constructed 6 81‐ha enclosures surrounded by 2.4‐m‐tall woven wire fence on each study site. The experimental design included 2 nutrition treatments and 3 deer densities in a factorial array, with study sites as blocks. Abundance targets for low, medium, and high deer densities in enclosures were 10 deer (equivalent to 13 deer/km ² ), 25 deer (31 deer/km ² ), and 40 deer (50 deer/km ² ), respectively. Each study site had 2 enclosures with each deer density. We provided deer in 1 enclosure at each density with a high‐quality pelleted supplement ad libitum , which we termed enhanced nutrition; deer in the other enclosure at each density had access to natural nutrition from the vegetation. We conducted camera surveys of deer in each enclosure twice per year and added or removed deer as needed to approximate the target densities. We maintained >50% of deer ear‐tagged for individual recognition. We maintained adult sex ratios of 1:1–1:1.5 (males:females) and a mix of young and older deer in enclosures. We used reconstruction, validated by comparison to known number of adult males, to make annual estimates of density for each enclosure in analysis of treatment effects. We explored the effect of deer density on diet composition, diet quality, and intake rate of tractable female deer released into low‐ and high‐density enclosures with natural nutrition on both study sites (4 total enclosures) between June 2009 and May 2011, 5 years after we established density treatments in enclosures. We used the bite count technique and followed 2–3 tractable deer/enclosure during foraging bouts across 4 seasons. Proportion of shrubs, forbs, mast, cacti, and subshrubs in deer diets did not differ ( P > 0.57) between deer density treatments. Percent grass in deer diets was higher ( P = 0.05) at high deer density but composed only 1.3 ± 0.3% (SE) of the diet. Digestible protein and metabolizable energy of diets were similar ( P > 0.45) between deer density treatments. Likewise, bite rate, bite size, and dry matter intake did not vary ( P > 0.45) with deer density. Unlike deer density, drought had dramatic ( P ≤ 0.10) effects on foraging of tractable deer. During drought conditions, the proportion of shrubs and flowers increased in deer diets, whereas forbs declined. Digestible protein was 31%, 53%, and 54% greater ( P = 0.06) during non‐drought than drought during autumn, winter, and spring, respectively. We studied the effects of enhanced nutrition on the composition and quality of tractable female deer diets between April 2007 and February 2009, 3 years after we established density treatments in enclosures. We also estimated the proportion of supplemental feed in deer diets. We used the 2 low‐density enclosures on each study site, 1 with enhanced nutrition and 1 with natural nutrition (4 total enclosures). We again used the bite count technique and 2–3 tractable deer living in each enclosure. We estimated proportion of pelleted feed in diets of tractable deer and non‐tractable deer using ratios of stable isotopes of carbon. Averaged across seasons and nutrition treatments, shrubs composed a majority of the vegetation portion of deer diets (44%), followed by mast (26%) and forbs (15%). Enhanced nutrition influenced the proportion of mast, cacti, and flowers in the diet, but the nature and magnitude of the effect varied by season and year. The trend was for deer in natural‐nutrition enclosures to eat more mast. We did not detect a statistical difference ( P = 0.15) in the proportion of shrubs in diets between natural and enhanced nutrition, but deer with enhanced nutrition consumed 7–24% more shrubs in 5 of 8 seasons. Deer in enhanced‐nutrition enclosures had greater ( P = 0.03) digestible protein in their overall diet than deer in natural‐nutrition enclosures. The effect of enhanced nutrition on metabolizable energy in overall diets varied by season and was greater ( P < 0.04) for enhanced‐nutrition deer during summer and autumn 2007 and winter 2008. In the enhanced‐nutrition treatment, supplemental feed averaged 47–80% of the diet of tractable deer. Of non‐tractable deer in all density treatments with enhanced nutrition, 97% ( n = 128 deer) ate supplemental feed. For non‐tractable deer averaged across density treatments, study sites, and years, percent supplemental feed in deer diets exceeded 70% for all sex and age groups. We determined if increasing deer density and enhanced nutrition resulted in a decline in preferred forbs and shrubs and an increase in plants less preferred by deer. We sampled all 12 enclosures via 20, 50‐m permanent transects in each enclosure. Percent canopy cover of preferred forbs was similar ( P = 0.13) among deer densities averaged across nutrition treatments and sampling years (low density: = 8%, SE range 6–10; medium density: 5%, 4–6; high density: 4%, 3–5; SE ranges are presented because SEs associated with backtransformed means are asymetrical). Averaged across deer densities, preferred forb canopy cover was similar between nutrition treatments in 2004; but by 2012 averaged 20 ± 17–23% in enhanced‐nutrition enclosures compared to 10 ± 8–13% in natural‐nutrition enclosures ( P = 0.107). Percent canopy cover of other forbs, preferred shrubs, other shrubs, and grasses, as well as Shannon's index, evenness, and species richness were similar ( P > 0.10) among deer densities, averaged across nutrition treatments and sampling years. We analyzed fawn:adult female ratios, growth rates of fawns and yearlings, and survival from 6 to 14 months of age and for adults >14 months of age. We assessed adult body mass and population growth rates (lambda apparent, λ APP ) to determine density and nutrition effects on deer populations in the research enclosures during 2004–2012. Fawn:adult female ratios declined ( P = 0.04) from low‐medium density to high density in natural‐nutrition enclosures but were not affected ( P = 0.48) by density in enhanced nutrition enclosures although, compared to natural nutrition, enhanced nutrition increased fawn:adult female ratios by 0.15 ± 0.12 fawns:adult female at low‐medium density and 0.44 ± 0.17 fawns:adult female at high density. Growth rate of fawns was not affected by deer density under natural or enhanced nutrition ( P > 0.17) but increased 0.03 ± 0.01 kg/day in enhanced‐nutrition enclosures compared to natural nutrition ( P < 0.01). Growth rate of yearlings was unaffected ( P > 0.71) by deer density, but growth rate increased for males in some years at some density levels in enhanced‐nutrition enclosures. Adult body mass declined in response to increasing deer density in natural‐nutrition enclosures for both adult males ( P < 0.01) and females ( P = 0.10). Enhanced nutrition increased male body mass, but female mass did not increase compared to natural nutrition. Survival of adult males was unaffected by deer density in natural‐ ( P = 0.59) or enhanced‐ ( P = 0.94) nutrition enclosures. Survival of adult females was greatest in medium‐density enclosures with natural nutrition but similar at low and high density ( P = 0.04). Enhanced nutrition increased survival of females ( P < 0.01) and marginally for males ( P = 0.11). Survival of fawns 6–14 months old was unaffected ( P > 0.35) by density in either natural‐ or enhanced‐nutrition treatments but was greater ( P = 0.04) under enhanced nutrition. Population growth rate declined ( P = 0.06) with increasing density in natural‐nutrition enclosures but not ( P = 0.55) in enhanced nutrition. Enhanced nutrition increased λ APP by 0.32. Under natural nutrition, we found only minor effects of deer density treatments on deer diet composition, nutritional intake, and plant communities. However, we found density‐dependent effects on fawn:adult female ratios, adult body mass, and population growth rate. In a follow‐up study, deer home ranges in our research enclosures declined with increasing deer density. We hypothesized that habitat quality varied among home ranges and contributed to density‐dependent responses. Variable precipitation had a greater influence on deer diets, vegetation composition, and population parameters than did deer density. Also, resistance to herbivory and low forage quality of the thornshrub vegetation of our study sites likely constrained density‐dependent behavior by deer. We posit that it is unlikely that, at our high‐density (50 deer/km ² ) and perhaps even medium‐density (31 deer/km ² ) levels, negative density dependence would occur without several wet years in close association. In the past century, this phenomenon has only happened once (1970s). Thus, density dependence would likely be difficult to detect in most years under natural nutrition in this region. Foraging by deer with enhanced nutrition did not result in a reduction in preferred plants in the vegetation community and had a protective effect on preferred forbs because ≤53% of deer diets consisted of vegetation. However, enhanced nutrition improved fitness of individual deer and deer populations, clearly demonstrating that nutrition is limiting for deer populations under natural conditions in western South Texas. © 2019 The Authors. Wildlife Monographs published by Wiley Periodicals, Inc. on behalf of The Wildlife Society.