The turnover of dental microwear texture: Testing the” last supper” effect in small mammals in a controlled feeding experiment

Highlights

• Turnover time of DMT is tested on two experimental diets.

• Subsequent change in DMT parameters is observed after switch on new diet.

• Between 16 and 24 days after switch, previous diet signal is overprinted.

• No immediate “last supper” effect visible, turnover needs at least two weeks.

Abstract

Dental microwear texture (DMT) analysis is commonly applied for dietary reconstruction of vertebrates. The temporal scale on which dietarily informative microscopic wear forms on enamel surfaces is crucial to infer dietary flexibility and seasonality. Microwear is assumed to form shortly before the individual's death, reflecting information pertaining to the last meals consumed (“last supper” effect). In primate feeding experiments, microwear features formed within hours, suggesting rates of turnover within one to two weeks. As DMT formation experiments testing the persistence of microwear three-dimensionally (textures) are still lacking, we test how quickly DMTs form and pre-existing ones are overwritten in a terminal feeding experiment with 72 rats. In two groups of 36, rats received either a standard pelleted diet or the same pelleted diet containing 4% loess, an aeolian, silt-sized sediment, for 24 consecutive days. Then 6 individuals from each group were sacrificed, while the rest (n = 30) were switched to the diet they had not received before. On day 1, 2, 4, 8, 16, and 24 after the diet switch, 5 of the remaining individuals were sacrificed, creating a cohort of n = 5 each for each time point. We applied DMT analysis on first and second upper molars. For upper second molars, rats show a subsequent change in DMT after the switch, with visible differences from day 2 on. On upper first molars, microwear textures were variable for individuals sacrificed directly after the initial 24-day feeding period, thus obscuring significant differences in diet-induced dental wear. We find turnover faster and more pronounced when switching from loess-containing to standard pellet as compared to the opposite switch. The trend for either decreasing or increasing parameter values after the diet switch approaches a plateau between 16 and 24 days for many DMT parameters, suggesting that, under these experimental conditions, the “last supper” effect needs at least two weeks to overwrite previous DMT patterns.

Introduction

Dental microwear and microwear texture analyses (DMTA) have been widely applied to reconstruct past and present diets in terrestrial mammals such as bovids (Ungar et al., 2007; Merceron et al., 2010; Schulz et al., 2010, Schulz et al., 2013; Scott, 2012; Winkler et al., 2013), carnivores (Ungar et al., 2010; DeSantis et al., 2012; DeSantis et al., 2013), primates (Calandra et al., 2012; Schulz-Kornas et al., 2019), hominins (Scott et al., 2005; Ungar et al., 2008), as well as mammaliaforms (Kalthoff et al., 2019) and recently also fish (Purnell and Darras, 2016), aquatic mammals (Purnell et al., 2017) and Lepidosauria (Winkler et al., 2019a; Bestwick et al., 2019). In these reconstructions, the temporal scale on which microwear is formed and overwritten is crucial for interpretation of intra-specific dietary flexibility, seasonality, environmental change and palaeodietary reconstruction. It further has implications for minimum length of feeding experiments where an overwriting of wear patterns is assumed (Hoffman et al., 2015).

For more than three decades, researchers have cited the “last supper” effect (Grine, 1986), which implies that microwear, and consequently microwear texture, reflects the last few meals of an individual (e.g., Solounias et al., 1994; Mainland, 1998; Grine et al., 2006; Joomun et al., 2008; Scott et al., 2009; Aiglstorfer et al., 2014; Rivals et al., 2015; Calandra et al., 2016; Martinez et al., 2016; Semprebon et al., 2016). However, in his original publication, Grine was only proposing that “…because microwear patterns can change as diet is altered (Walker et al., 1978; Covert and Kay, 1981; Teaford and Oyen, 1986), the items that were masticated by an individual over a period of time just prior to its death will have a profound effect upon interpretations of its dietary habits from details of tooth wear…”. The studies he based his hypothesis on documented formation of in-vivo microwear in vervet monkeys every six weeks (Teaford and Oyen, 1986), in opossums fed cat food or cat food with added plant fibre or chitin after 90 days, and additionally ground pumice after the last 30 days (Covert and Kay, 1981), and in wild populations of hyrax on single individuals that died during the wet or dry season (Walker et al., 1978). None of these original studies allows the conclusion that microwear actually forms within only days, because the shortest experimental duration for which microwear was measured after a dietary switch was six weeks.

Teaford and Oyen (1989) were the first to experimentally address the question of how quickly microwear features, i.e. pits and scratches, form on the enamel surface of teeth. They took dental impressions of vervet monkeys raised on hard (Purina Monkey Chow diet no. 5038 supplemented with apples) or soft food (water softened Monkey Chow and pureed fruit cocktail) three times in four days, establishing a baseline and evaluating 24 h and 72 h post-baseline. After 72 h, they found up to 26% new wear features on enamel shearing facets, and up to 80% new features on grinding facets, thus interpolating that old microwear features could be replaced completely in 1–2 weeks (Teaford and Oyen, 1989). Studies of dental patients (Teaford and Tylenda, 1991) and Costa Rican howler monkeys (Teaford and Glander, 1991, Teaford and Glander, 1996) demonstrated that microwear formation on a normal diet could vary drastically between species and settings. In a single-subject study on a human, Teaford and Lytle (1996) found 90% new features on the occlusal surface after one week of feeding on a diet containing mineral residues from sandstone grinding of corn. Romero et al. (2012) found buccal microwear in human subjects to quickly respond to a more abrasive (stone-ground) diet, with 22% new features formed per week compared to ~1.9–2.8% per week under a Mediterranean diet. In more recent studies of occlusal microwear in non-human primates (Teaford et al., 2017, Teaford et al., 2020), new microwear features were reported to appear after a single feeding bout with brazil nuts. The number of new features positively correlated with the number of brazil nuts consumed in that bout, but the percentage of these new features as compared to the overall microwear features was very small (0–6%), suggesting that a time period of more than a single meal is required to consistently change a microwear surface. Overall, Teaford et al. (2013) concluded that the formation of microwear patterns is “essentially a complex summary of past events, often with multiple signals superimposed on each other” and that the “last supper” effect “is surely situation-specific and depends on a wide range of factors” (Teaford et al., 2020).

These studies are the only ones systematically addressing the rates of dental microwear formation, all using scanning electron microscopy to visualize microwear features. The conclusions obtained under controlled experimental conditions are, however, limited to non-human primates up to 72 h after a diet change and a single human subject after one week, and none of these studies actually addressed the question of how quickly dental microwear patterns (now most-commonly assessed via various forms of DMTA) could change.

To address the question of how fast a pre-existing DMT is completely overwritten, we performed a terminal feeding experiment, using rats (Rattus norvegicus forma domestica) as a model organism. In order to determine how fast pre-existing dental wear patterns were overwritten, we tested both the turnover time and dietary abrasiveness when foods were switched from a low abrasive (standard rat pellet) to a higher abrasive diet (standard rat pellet with 4% loess added) and vice versa. In contrast to previous studies that used microscopic inspection of type and abundance of wear features, we employ DMTA to characterise the complete surface topography of small enamel areas (Ungar et al., 2003; Scott et al., 2005; Calandra et al., 2012; Schulz et al., 2013; Winkler et al., 2019a, Winkler et al., 2019b).