On the Effects of Acid Pre-treatment on the Elemental and Isotopic Composition of Lightly- and Heavily-calcified Marine Invertebrates

Carbonate removal using acids is a common practice in ecological studies. The effects, however, of acid pre-treatment on the elemental and isotopic composition of marine invertebrates as well as how these effects vary according to species’ carbonate content is little known. We examined the effects of acid pre-treatment on the elemental (%C, %N, C:N ratio (%C:%N)) and isotopic composition (δ13C, δ15N) of 28 lightly- and heavily-calcified species from Cnidaria, Mollusca, Arthropoda, Bryozoa, Echinodermata and Chordata. The present study showed that acid pre-treatment modified the elemental and isotopic composition of lightly- and heavily-calcified marine invertebrates. The shifts were clearly seen as a decrease in the %C and δ13C of heavily-calcified species while we did not detect a clear pattern for %N and δ15N (in both lightly- and heavily calcified species). Apart from carbonates, acid pre-treatment caused also the loss of organic compounds, thus confounding the interpretation of carbonate proxy (CP) -a widely used proxy for carbonate content. We recommend the use of CP solely with heavily-calcified species. For the first time it was shown that the use of δ15N values from acidified samples can introduce substantial bias in our perception about the number of trophic levels, the distribution of species and distribution of biomass across the trophic levels in a community. We have uncovered and elucidated previously unknown aspects and highlighted the challenge posed when predicting shifts in elemental and isotopic composition of species following acid pre-treatment. The present findings should be considered in future studies using acid pre-treatment as they can contribute to the optimum use of samples while avoiding bias in the interpretation of findings.


2
For the first time it was shown that the use of δ 15 Ν values from acidified samples can 27 introduce substantial bias in our perception about the number of trophic levels, the distribution 28 of species and distribution of biomass across the trophic levels in a community.

29
We have uncovered ravelledand elucidated previously unknown aspects and highlightedting

43
The elemental and isotopic composition of organisms' organic compounds can provide 44 ecologists with useful information about food-web structure and cycling of organic matter.

91
The present study is the first one to investigateing the effects of acid pre-treatment on the 92 elemental and isotopic composition of marine invertebrates from a large range of carbonate 93 content [from lightly (e.g. arthropods) to heavily-calcified (e.g. echinoderms)] and the possible 94 consequences of these effects on the interpretation of community trophic structure. It was 95 hypothesized that carbonate removal would cause the depletion of %C and δ 13 C values -96 especially in heavily-calcified organisms like echinoderms and calcified bryozoans-while such an 97 effect was not expected for %N and δ 15 N. In addition, it was hypothesized that no differences in  and nitrogen isotope ratios (Bosley & Wainright 1999;Post 2002;Sarakinos et al. 2002;Fanelli 120 et al. 2010;Ruiz-Cooley et al. 2011;Lau et al. 2012;Rennie et al. 2012;Liu et al. 2013) we 121 acknowledge that the δ 13 C and δ 15 N values of fauna from cold-water coral reefs may have been 122 distorted due to the chosen method of sample preservation. The possible distortion of stable 123 isotope ratios due to preservation in formalin could introduce some uncertainty in the 124 examination of a community's trophic structure (i.e. carbon sources, species' trophic level); 125 however, it should be mentioned that a) several studies have shown that the effects of formalin 126 preservation on δ 15 Ν values were minor compared to a commonly-used trophic fractionation 127 factor (i.e. +3. 4‰, DeNiro and Epstein 1981;Post 2002) enabling thus the allocation of species to trophic levels (Fanelli et al. 2010) and b) a confounding effect on our findings should not be 6 expected since both acidified and non-acidified subsamples have been preserved in formalin.

Sample preparation for stable isotope analysis
Samples fixed in formalin were dried at 60 o C while frozen samples were freeze-dried (0.041 difficulty into removinge all small shell pieces. Similarly, the shell of Pagurus sp. was removed 161 before grinding. Grinding was carried out using the TissueLyser II (Qiagen) or a mortar and pestle 162 according to the size of the organisms. Following grinding, each species was subjected to a 163 preliminary analysis of their C and N content (as % of dry mass) in order to assess the optimal 164 amount of dry mass for dual (δ 13 C and δ 15 N) stable isotope ratios analysis. Each sample of macro-165 and megafauna was divided in two subsamples i.e. subsamples to be acidified or not.

166
Subsamples not to be acidified were placed in tin cups while subsamples to be acidified were 167 placed in silver cups (5x8 mm, Elemental Microanalysis UK). Carbonate removal was carried out 168 through the sequential addition of 15µl of (1M) hydrochloric acid directly in the silver cups using 169 a micropipette; the cessation of the effervescence was used as the criterion that carbonates had 170 been removed (Vafeiadou et al. 2013). The total volume of 1M hydrochloric acid added in each 171 silver cup was recorded (Table S2). Both acidified and non-acidified samples were dried at 60°C 172 overnight (e.g. Carabel et al. 2006;Jaschinski et al. 2008;Serrano et al. 2008); no washing with 173 distilled water was carried out (Mateo et al. 2008). Overnight drying and no washing with 174 distilled water were also performed for the non-acidified subsamples.

203
Statistical differences between acidified and non-acidified subsamples for the parameters %C, 204 %N, C:N ratio, δ 13 C, δ 15 N were examined. The comparisons between acidified and non-acidified 9 subsamples for the elemental and isotopic composition were carried out with the paired Student's t-test or non-parametric Wilcoxon signed rank-test (significance level p<0.05).

207
Beforehand, the normality of the distributions was checked with the Shapiro-Wilk test and 208 equality of variances with the F-test. The correlations between the CP and i) δ 13 Cno acid-δ 13 Cacid, 209 ii) δ 15 Νno acid-δ 15 Νacid, were examined with Spearman's correlation (rs). Examination of 210 differences and correlations were carried out at the statistical analysis environment R. Only 211 species with at least three replicates were included in the statistical analysis.

214
The effects of carbonate removal were further examined through the comparison of the trophic The benthic community that was examined is the recently-described association of the cold-   Table 1). The highest decreases in %C were found in 246 echinoderms (-9.4±1.6, P<0.001) and bryozoans (-5.0±2.4, P<0.01) (i.e. in organisms with 247 carbonate proxy values higher than +0.5). In a previous study by Serrano et al. (2008) on beach 248 arthropods the decrease in %C was attributed to the loss of both inorganic (i.e. carbonates) and 249 organic carbon after carbonate removal. Loss of organic carbon may have also played a role in 250 the %C decrease in species with low carbonate content and especially in the isopod Janira 251 maculosa and the chiton cf Tonicella marmorea where the decrease in %C was comparable to 252 that found in echinoderms (Table 1). For those species with high carbonate content (e.g. that the significant decrease (P<0.01) in %C was driven by the removal of carbonates.

255
As regards %N, echinoderms were the only group where a statistically-significant decrease

267
In 67% of the species, there was a statistically-significant decrease in the C:N ratio, which 268 shows a proportionally-higher loss of carbon than nitrogen (Table 1) as well as the differences 269 between carbon and nitrogen as regards their response to carbonate removal using hydrochloric 270 acid. The decrease in the C:N ratio was higher in heavily-calcified species i.e. calcified bryozoans 271 and echinoderms (-3.6%) than lightly-calcified species i.e. cnidarians and arthropods (-0.9%).

272
The proportionally higher loss of carbon than nitrogen in the heavily-calcified species played a

280
In previous studies negative CP values were attributed to a proportionally higher loss of %N than increase in %C and %N in acidified subsamples through loss of molecules comparatively low in     Table 2 for 358 details).

359
The findings described above provided evidence that the use of δ 15 Ν values from acidified

525
Applying stable isotopes to examine food-web structure: an overview of analytical tools.