Mining of minerals has been observed in adult Mohave Desert Tortoises (Gopherus agassizii; Marlowe and Tollestrup 1982) and possibly in juvenile Sonoran Desert Tortoises (Gopherus morafkai; Stitt and Davis 2003). However, little is known about the timing of consumption or composition of putatively mineral-rich rocks or soil ingested by adult Sonoran Desert Tortoises. Hazard et al. (2010) suggested that consumption of calcium-rich materials (stones, caliche soil) by female tortoises is consistent with the notion that these tortoises require additional calcium during or following egg production. Herein we describe observations on the temporal dynamics of calcium-rich caliche consumption by female Sonoran Desert Tortoises supporting this hypothesis.

Since 2011, we have been conducting a long-term radio-tracking study of Sonoran Desert Tortoises on the eastern slopes of the southeastern Union Hills on the northern edge of the Phoenix Metropolitan area. While recording feeding behavior over the past 8 yrs, we have observed consumption of small bits of caliche and other mineral deposits by adult female Sonoran Desert Tortoises (straight-line carapace length [SLCL] > 220 mm). Although consumption of calcium-rich items, including bone, rocks, and soil layers, has been documented in desert tortoises of the Southwest (see reviews in Walde et al. 2007; Lovich et al. 2018), our observations suggest that female Sonoran Desert Tortoises only consume caliche and other mineral deposits during a relatively brief period in late spring and very early summer, when conditions are especially harsh with respect to water balance (Averill-Murray 2002; Van Devender 2002). The observation that female Sonoran Desert Tortoises consume hard, dry bits of caliche when they are otherwise largely inactive (late May/ June), and have had no access to free-standing water for months, suggests that these minerals are indeed critical at this time.

Field Observations. — Our Union Hills study site (∼ 33.73°N, 112.06°W) is composed of a relatively low-elevation series of hills in a region of transition from creosote (Larrea tridentata)–bursage (Ambrosia deltoidea) flats to saguaro (Carnegiea gigantea)–palo verde (Parkinsonia microphylla)-dominated uplands. The hills rise to approximately 650 m elevation from a surrounding plain of about 350 m; geologically they comprise metavolcanic rocks with basaltic protoliths and various granitic rocks in lesser quantity. Within this site, tortoises are rarely encountered away from the slopes and the incised arroyos draining the hills, which are dominated by plants associated with the Arizona Upland Subdivision of the Sonoran Desert (Turner and Brown 1982). The arroyos have numerous exposed caliche formations into which tortoises and other organisms have excavated burrows used by tortoises during both winter and summer. Given the limited availability of large rock outcrops, tortoises at our site did not typically take refuge under rock overhangs, as is commonly observed in other Sonoran Desert populations (Barrett 1990; Averill-Murray et al. 2002). We consistently observed both male and female Sonoran Desert Tortoises using deep, north-facing caliche tunnels during May and June, the hot and dry portion of late spring and early summer in the Sonoran Desert (Sullivan et al. 2016).

Over the past 8 yrs we radio tracked 15–25 adult Sonoran Desert Tortoises each year, gathering detailed observations on annual variation in space use and activity. When radio-tracking tortoises, we opportunistically recorded feeding behavior and collected data on hundreds of feeding episodes. We noted evidence of feeding during every month of the year, but we obtained detailed recordings (video, photographs) of feeding sequences for subsequent analysis during 9 mo (February through October). Over 8 yrs of observations on feeding, we occasionally observed consumption of small bits of caliche, or what we term “caliche fragments,” encountered apparently at random along the wash bed and adjacent slopes.

To assess the chemical composition of caliche fragments consumed by tortoises, we gathered samples on 3 routes (2–3-m pathways) near caliche caves along which 2 female Sonoran Desert Tortoises (female 22 = 2 routes, female 51 = 1 route) were observed consuming small white fragments or actual portions of a large caliche formation on at least 2 occasions. Using a meter stick laid directly over the pathway taken by the female after she was initially observed ingesting a fragment as a guide, we picked up 5 caliche fragments (one every 200 mm) approximately equal to the dimensions in size and color to those we observed consumed by female Sonoran Desert Tortoises previously (judged by eye). Samples were combined for each female Sonoran Desert Tortoise. We also collected 5 nonwhite, parent rock samples over the same distance (i.e., one every 200 mm) to compare with the caliche for the 2 routes taken by female 22; female 51 consumed caliche from a route lacking parent rock material at the surface along her pathway. The rocks were crushed into a powder and approximately 5–10 g of the powder was analyzed for its elemental composition by a Thermo Scientific Quant'X X-ray Fluorescence (XRF) spectrometer. This instrument irradiated samples with x-rays and then measured the subsequent emission of x-rays from the sample to identify elements. This instrument was able to quantify the concentrations of elements with atomic numbers greater than sodium. The results were expressed as percentage of the element in the sample.

Results and Discussion. — When ingesting caliche fragments, female Sonoran Desert Tortoises walked slowly forward, touching their snout to the ground regularly while swinging their head from side to side, similar to typical foraging behavior (Fig. 1). Immediately prior to consuming caliche fragments, female Sonoran Desert Tortoises repeatedly pressed their snout against an individual caliche fragment, as if ascertaining the suitability of the fragment for consumption (Fig. 1). Often, a fragment was passed over after a few apparent (closed mouth) sampling efforts, but others were grasped and bitten, with audible scraping and crunching sounds produced (Fig. 1). On multiple occasions pea-sized bits were consumed after 5–10 sec of mastication; female Sonoran Desert Tortoises never appeared to swallow the small fragment whole. On a few occasions, female Sonoran Desert Tortoises consumed bits of dry grass and other vegetation over a 5–10-min period of walking slowly forward, with repeated consumption of caliche fragments interspersed among those events. On one occasion, a female tortoise sampled dry, white powder lining the surface of a small (1-m-diameter) depression in a wash where water regularly evaporated, leaving behind the salt residue. While pressing her snout against the substrate she repeatedly grasped and bit small rocks coated with white mineral deposits (Fig. 2).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Consumption of caliche fragments or catchment minerals occurred on 10 occasions between 26 May and 27 June, over 5 yrs (2014–2018), by 4 different females. Over these same 5 yrs, we recorded detailed observations on 190 feeding bouts (plant material only) between 1 February and 31 October by 10 females and 6 males. Because foraging activity in tortoises may vary with rainfall and temperature (e.g., Duda et al. 1999; Sullivan et al. 2016), a statistical assessment of a seasonal bias in caliche consumption independent of feeding required that we control for annual variation in the distribution of foraging bouts. We used the observed distribution of feeding bouts (monthly) during the 5 yrs when we observed caliche consumption to evaluate the annual distribution of caliche consumption events. We grouped number of foraging and caliche consumption bouts observed into 3 seasonal periods of activity as noted by Sullivan et al. (2016): spring (February–April, n = 48 feeding bouts, 0 caliche consumption events), hot/dry spring (May–June, n = 31 feeding bouts, 10 caliche), and monsoon/fall (July–October, n = 113 feeding bouts, 0 caliche). By contrast to feeding events, caliche consumption observations were restricted to the latter half of the hot/dry (n = 10) period exclusively (Fisher's 2 × 3, p < 0.0001; 4 of 10 females, 0 of 6 males). All but one of the caliche consumption events (n = 9) were within 10 m of the female's deep (> 1 m) refuge tunnel along a wash bank; the lone exception was a female Sonoran Desert Tortoise that was some distance from her refuge when consuming the catchment minerals (n = 1; 40 m from refuge). No males were observed ingesting caliche fragments.

Our analysis indicates that by ingesting caliche fragments near refuges, female Sonoran Desert Tortoises obtain considerable calcium (Table 1). The average calcium concentrations in the samples associated with 2 female Sonoran Desert Tortoises were 22.1% and 25.6%, which contrasts with parent rock that averaged 1.14% calcium. Caliche is generally considered a soft, carbonate rock that forms cement with sand, gravel, and mud (Schlesinger 1982); interestingly, our caliche samples lacked the alkali elements of sodium and potassium. These elements are water soluble when paired with carbonates and therefore less likely to condense from an aqueous solution. The parent rock, granitic in appearance and apparently avoided by our female Sonoran Desert Tortoises, was richer in aluminum and silicates. Although some have suggested that geophagy might aid in the maceration of ingested plant materials (Sokol 1971; Kramer 1973), the acidic foregut of desert tortoises (Barboza 1995) would likely render easily decomposed caliche fragments ineffective in this capacity. The parent rock, ignored by the female Sonoran Desert Tortoises we observed, would have been a more suitable choice under this alternative functional hypothesis. Table 1 presents the most abundant and regularly detected elements in the samples, but sulfur, vanadium, zinc, rubidium, strontium, yttrium, zirconium, and barium were also detected in one or more of the caliche samples in trace amounts.

Table 1. Elemental concentrations (%, mean ± SD, n = 3) for samples of caliche fragments associated with areas used by Sonoran Desert Tortoise females 22 and 51 as well as typical rocks associated with Sonoran Desert Tortoise female 22. The intermediate rock samples showed some caliche encrusting the rocks, yielding intermediate calcium levels between caliche and native rock samples. ND = not detected.

View Fullscreen

Esque and Peters (1994) described female Mohave Desert Tortoises in southwestern Utah ingesting small white stones, which they suggested were composed of calcium carbonate. Female Mohave Desert Tortoises will move significant distances to mine minerals in soil (Marlowe and Tollestrup 1982), and they will consume bones of various vertebrates (Walde et al. 2007; Lovich et al. 2018). Our observations reveal that, at least within a caliche-rich wash system, female Sonoran Desert Tortoises may not need to move far to obtain calcium (but not necessarily phosphorus) sources and that these consumption events are clustered close to the reproductive period typically followed by egg laying. Lovich et al. (2017) documented that for Sonoran Desert Tortoises, the average date of first appearance of shelled eggs within reproductive tracts is 6 June and the average date of last appearance is 26 June, which accords well with our dates of caliche consumption of 26 May to 27 June. Given that Lovich et al. (2017) detected shelled eggs somewhat earlier than most of our field observations, it may be that female Sonoran Desert Tortoises consume caliche to replenish calcium expended immediately prior during egg production.

Our observations are limited by the relatively small sample size of observations (n = 10) and even lower number of individual female Sonoran Desert Tortoises (n = 4), but are suggestive of a relationship between egg-laying and ingestion of minerals, at least for a population of Sonoran Desert Tortoises in a caliche-lined wash. Some have argued that the herbivorous diet of desert tortoises should provide considerable calcium, rendering active consumption unnecessary (e.g., Oftedal 2002). Nonetheless, the consistency with which female Sonoran Desert Tortoises (this study) and Mohave Desert Tortoises (Marlowe and Tollestrup 1982; Esque and Peters 1994; also see Lovich et al. 2018) actively seek out calcium-rich materials provides compelling evidence of a linkage between mining and egg production. Interestingly, juvenile Sonoran Desert Tortoises have been observed mining during the monsoon (July–August; A. Owens, pers. comm., July 2018); of 2 juveniles (SLCL < 180 mm) in our study population, the female juvenile was observed scraping minerals with its forelimbs while inside a large caliche refuge (cave), briefly during July, but we were unable to confirm consumption. Unlike the population of Mohave Desert Tortoises evaluated by Esque and Peters (1994), in which very few tortoises engaged in mineral consumption when feeding (∼ 1%), in our population of Sonoran Desert Tortoises 40% of females under observation for feeding bouts ingested caliche fragments on at least 1 occasion. Additional study will be necessary to more fully evaluate our hypothesis of a temporal link between caliche consumption and egg production in Sonoran Desert Tortoises.