Preservation effects on wet weight, dry weight, and ash-free dry weight biomass estimates of four common estuarine macro-invertebrates: no difference between ethanol and formalin
© Springer-Verlag and AWI 2005
Received: 5 July 2004
Accepted: 8 April 2005
Published: 14 May 2005
Two preservative treatments traditionally used in aquatic sciences, formalin (4%) and ethanol (70%), were compared for their effects on biomass estimations. The effects of both preservatives on wet weight, dry weight, and ash-free dry weight were determined for samples preserved for 10, 21, and 90 days. The effects were studied in four different macrofauna species commonly found in German estuaries: Heteromastus filiformis (Capitellidae, Polychaeta), Hediste diversicolor (Nereididae, Polychaeta), Corophium sp. (Amphipoda, Crustaceae), and Gammarus spp. (Amphipoda, Crustacea). The biomass estimates of preserved samples were compared with those of unpreserved samples. In all four species the loss in wet weight, dry weight, and ash-free dry weight was most pronounced within the first 10 days, and an additional weight loss was recorded between days 10 and 21. However, there was no further loss in weight for samples kept for as long as 90 days in the preservatives. In general, crustaceans exhibited higher weight loss than polychaetes, and smaller species (H. filiformis and Corophium sp.) showed higher weight loss and a higher variability than larger species. As our main result, significant differences between the two preservative treatments did never occur. Our results contradict some earlier investigations on this matter where formalin has been reported to be superior to alcoholic preservatives because weight loss was less pronounced than in ethanol. Factors affecting biomass estimates are discussed and we conclude that, for the macrofauna groups tested, the use of the toxic formalin solution is not justified when the major intent is biomass estimation.
In aquatic studies biomass estimates are commonly performed on preserved material because extensive sampling often requires preservation in the field before determination and analysis can be carried out. In marine studies the most common preservative for macrofauna is formalin (4% final concentration), while in limnology the use of ethanol (70–80% final concentration) prevails. Formaldehyde is a colorless gas highly soluble in water. It is commercially sold as saturated aqueous solution with concentration of about 37%, conventionally called 100% formalin. Working solutions are normally buffered with sodium tetraborate (Borax), hexamethylene tetramine (e.g. Ruhmor 1990), or phosphate in order to prevent the formation of formic acid, which can substantially degrade the tissue quality.
The health hazard involved with the use of formalin has been covered extensively (e.g. Black and Dodson 2003) and has been recognized by most scientists today. In brief, as a rule of thumb, the slightest odor of formalin in the air indicates concentrations as small as 0.6 ppm (AIHA 1989), which can already cause acute exposure effects such as irritation of eyes, nose, and throat in some individuals. Higher concentrations, in the range of 3–5 ppm, will cause eyes to water (an experience most biologists have made at least once during their career), while concentrations above approximately 25 ppm can cause severe injuries of the respiratory tract. Moreover, short-term contact with formalin solutions can result in severe skin irritation and the individual can be sensitized to formalin resulting in a allergic reaction during future contacts. In addition, the carcinogenic nature of formaldehyde has to be taken into account. It is a risk, which might even exist when individuals are exposed to formaldehyde concentrations below the 0.6 ppm olfactorial threshold. As a consequence, the use of formalin should be reduced as much as possible.
Despite the considerable human health concerns involved with its use, many marine biologists continue to use formalin, because it has become a standard procedure for preserving fauna samples (e.g. Ruhmor 1990) and is recommended by many official sources such as the Baltic Marine Environment Protection Commission (Helsinki Commission, HELCOM) (e.g. HELCOM 1988), the International Council for the Exploration of the Sea (ICES 1994; ICES 1996), and the Convention for the Protection of the Marine Environment of the North-East Atlantic (OSPAR Convention) (OSPARCOM 1997). In some cases formalin is a reasonable fixative especially when animals are collected for museum collections (one aspect of fixation with formalin is that it stiffens the body while ethanol often results in rather squashy specimens). The price difference between the two preservatives also plays a considerable role. In Germany, for example, some laboratory retailers charge almost twice as much for ethanol (96%, industrial grade) as for the same amount of formaldehyde solution (>35%). In countries where alcohol is even more prohibitively priced this difference can be even higher. However, a different perspective might be achieved, if the additional costs for the increasing safety requirements (protective close, respirators, fume hoods, etc.) and if differences in hazardous waste fees are taken into account.
In most studies samples are collected in order to gain information about species compositions and abundance, a task similarly achievable when formalin is replaced by ethanol. However, some previous studies have reported that for the assessment of biomass, an important parameter describing standing crop in benthic communities, ethanol is not recommended because weight loss is higher than in formalin preserved specimens (Howmiller 1972; Dermott and Paterson 1973; Landahl and Nagell 1978; Leuwen et al. 1985), while other studies could not confirm this observation (Mason et al. 1983; Gaston et al. 1996). Moreover, in some cases ethanol-fixed specimens continued to lose weight over time (e.g. Dermott and Paterson 1973) – a fact which would make biomass estimates unreliable, and most likely resulted in the recommendation of formalin as the fixative of choice for marine benthos.
In this study, we (1) established conversion factors for wet weight (WW), dry weight (DW), and ash-free dry weight (AFDW) for four common macrofauna species (Heteromastus filiformis, Hediste diversicolor, Corophium sp., and Gammarus spp.), and (2) tested if there are significant differences between biomass estimates when samples are preserved in either 4% formalin or 70% ethanol. In addition the weight loss is described over a time period of 90 days.
Material and methods
Measurements were made on material collected in the Weser estuary, Germany. This estuary is located at the southern coast of the North Sea and discharges the river Weser (watershed area 46,306 km2) into the Wadden Sea. The estuary mouth is located close to the city of Bremerhaven and is characterized by diurnal tides (mean tidal range at the Bremerhaven tidal gage is about 3.8 m) and mean water temperatures range from approximately 5°C in the winter time to 20°C in summer, while salinity can fluctuate between about 8 to 15‰ depending on season, river run-off, and tidal cycle.
For the present investigation four species from the two most abundant macrofauna taxa, polychaetes, and crustaceans were chosen. This species selection was guided by their relative abundances in the estuary and the desire to compare larger and smaller species from the same taxa in order to cover possible differences in conservation effects on biomass estimates between larger and smaller animals. Heteromastus filiformis (Claparede 1864) was chosen as a smaller polychaete (length up to 100 mm), while Hediste diversicolor (O.F. Müller 1776) was selected as a representative of a larger polychaete (length up to 200 mm). In crustaceans we chose Corophium sp. Latreille, 1806 as a small (length up to 10 mm) and Gammarus spp. Fabricius, 1775 (Gammarus salinus Spooner, 1947 and G. zaddachi Sexton, 1912) as a larger species (length up to 20 mm). Most of these species are inhabitants of eulitoral muddy sediments and were collected at low tide on a tidal flat close to the city limit of Bremerhaven (53°33.63′N, 8°31.34′E, WGS 84, Mercator projection, ellipsoid). Only the gammarids, were collected from artificial hard substrate periphyton (surface marker buoy No. 33 Fedderwarder Ost, 53°41.72′N, 8°20.06′E, WGS 84, Mercator projection, ellipsoid). Samples were collected and sorted directly at the sites, then flushed with water to remove any adhering sediment or detritus. Living animals were brought to the laboratory, were they were divided into replicate subsamples prior to fresh weight (WW of unpreserved samples, FW) estimation (see below). Due to the large differences in weights between the different taxa and sizes, cohorts of several individuals in each replicate where used, each made up of 10 (H. filiformis), 5 (H. diversicolor), 100 (Corophium sp.), and 10 individuals (Gammarus spp.). This procedure helped to increase precision of weight estimates. Altogether, 84 subsamples (21 per macrofauna species) were prepared and randomly assigned to one of the treatments (7 per species, see below). However, 10 days treatments in H. diversicolor were excluded from the analysis due to extreme variabilities in weight estimates.
In one treatment of each group, weight estimates (FW, DW, and AFDW) were measured using unpreserved samples. For the remaining six treatments the FW was estimated by blotting the live animals on tissue paper for 1 min (cf. Dermott and Paterson 1973) prior to weighing to the nearest ±0.1 mg. Then the samples were preserved either in borax buffered formalin (4%) or ethanol (70%), were they were kept for 10, 21, or 90 days at about 15°C in the dark. After that, samples were blotted for 1 min on tissue paper and the WW was estimated to the nearest ±0.1 mg. Next the DW of the samples was measured. Samples were placed for 24 h in an oven at 100°C, cooled down to room temperature in a desiccator (3–4 h), and then weighed. To determine the ash content of the samples, these were placed into a muffle furnace at 550°C for 1 h, cooled down to room temperature (6 h) in a desiccator, and then weighed. The AFDW was calculated by subtracting the ash-weight from the DW.
The WW was calculated as percentages of the initial FW, while DW and AFDW are given as percentages of the respective weights of unpreserved samples. In order to identify significant differences between treatments we calculated the 95% confidence intervals (CI) using Tukey’s honestly significant difference method (Tukey’s HSD or T-method, see Sokal and Rohlf 1998). Non-overlap in 95% CI corresponds to significant differences at α = 0.05 between samples. Statistical power (type II error probability, 1-β) was calculated using algorithms for balanced analysis of variance tests (see Crawley 2002). High power values (i.e. >80%, also commonly expressed as 0.8) indicate high probabilities to find significant differences when they exist. All calculations and figure compilations were performed using the freely available R computer language, version 1.7 (Ihaka and Gentleman 1996, for information and downloads see also http://www.r-project.org).
DW and AFDW in percent of the initial FW of unpreserved samples for four different macrofauna groups (means ± SD)
DW (% of FW)
AFDW (% of FW)
AFDW (% of DW)
Means of WW, DW, and AFDW (in percent of the corresponding weight estimates from unpreserved samples ±95% CI; DW and AFDW are also given in percent of WW ±95% CI) in four macrofauna groups preserved in formalin (4%) or ethanol (70%) for 10, 21, and 90 days, respectively
Ethanol (70 %)
Formalin (4 %)
WW (% of WW from unpreserved samples = FW)
DW (% of DW from unpreserved samples)
DW (% of WW)
AFDW (% of AFDW from unpreserved samples)
AFDW (% of WW)
In this study, we established conversion factors for biomass estimates for four common estuarine macrobenthos species preserved either in formalin (4% final concentration) or ethanol (70% final concentration) and demonstrated that the choice of the preservative is irrelevant for biomass estimation. In other words, ethanol preserved specimen did not lose more weight than specimens preserved with formalin. Nevertheless, both preservation methods resulted in a weight loss, which was most pronounced within the first 21 days. After 21 days, however, weight had stabilized and only small differences have been observed between samples kept for 21 and 90 days in fixative solutions. With our species selection we covered two common macrobenthos taxa, polychaetes, and crustaceans, both with a small and a large species. Smaller species showed greater weight loss (compared to unpreserved specimens) and exhibited higher variabilities than large ones. Though we cannot be sure that our conversion factors are entirely applicable to other related species within the same size category, they probably will not deviate too much.
In general, our data of biomass estimates of unpreserved samples are well in range compared with available data from the literature. A comprehensive compilation of some biomass conversion factors including those from unpublished sources and gray zone literature can be found for example in Rumohr et al. (1987). Therein, DW values (in persent of WW values) are given ranging from 12.9 to 29.9% (for Corophium) and from 15.4 to 26.4% (for Gammarus). Our results (14.3% for Corophium and 26.1% for Gammarus) are well placed within these ranges.
Previous studies on organisms collected from fresh water habitats (mainly chironomids) have reported higher weight losses of WW, DW or AFDW (e.g. Howmiller 1972; Dermott and Paterson 1973; Landahl and Nagell 1978; Leuwen et al. 1985) as well as higher losses of carbon content (Salonen and Sarvala 1985) in samples fixed with ethanol in comparison to formalin fixed specimens. Nevertheless, in some cases mass gain has been reported for ethanol fixed samples (e.g. Wiederholm and Eriksson 1977). However, our results for common marine/brackish macrofauna species do not support these findings. On the contrary, in most cases we found lower masses, though not significant, in samples preserved with formalin. Similar findings have been reported by Mason et al. (1983) and Gaston et al. (1996) for marine macroinvertebrates. One reason for these differing findings might be that insect larvae contain higher amounts of lipids in preparation for metamorphosis (e.g. Cavaletto et al. 2003), a substance easily soluble in alcoholic solvents. Marine invertebrates, in contrast, accumulate much less proteins and rarely have vast stores of lipids, except when gravid (Gaston et al. 1996). Moreover, our own observations on fixation practice with ethanol in freshwater species showed that fixation with highly concentrated ethanol preservatives (70–90%) can often result in the fast increases in body volume of chironomides causing cracks at the abdomen or head capsule with the consequence of losses in body fluids. It is evident that this will distort biomass estimates.
Sample handling is a great source of error for biomass estimation. For example: the removal of externally adhering fluids prior to weight estimation is a necessary step in WW estimation and different methods have been applied with different levels of success. Centrifugation of specimens (e.g. Howmiller 1972), for example, is less likely to produce reliable results because this most likely results in the loss of fluid from body tissues of preserved specimens (Wiederholm and Eriksson 1977) which, especially in cases of repetitive WW estimation, easily results in assumed weigh losses; intensity and duration of centrifugation most likely influence the result further (Mason et al. 1983). Alternatively, blotting samples with filter paper for defined time periods has been shown to be the most appropriate procedure although, the time period samples are blotted can significantly influence weight estimates (Dermott and Paterson 1973): one min of blotting has been shown to be most efficient and least error-prone. However, blotting followed by “air drying” for undefined time periods (Howmiller 1972; Hamilton and Kingston 1985) are not recommended, especially when comparisons of preservatives with different volatilities are intended (e.g. Hamilton and Kingston 1985, alcohol versus formalin). About 5–10 min of air exposure can induce WW losses of up to 25% (Jonasson 1972). Likewise, drying temperature is known to influence biomass estimation considerably. Higher weight losses have been reported when drying temperature in preparation of DW estimation exceeded 160°C, while temperatures below 500°C have proven to result in inhomogeneous AFDW estimates (Mason et al. 1983). Similarly important seem to be storage conditions. Leuwen et al. (1985) reported significant lower DW and AFDW looses for the gastropod Radix peregra when samples were stored in the dark.
Specimen condition, which can be substantially impaired by sample treatment (excessive sieving), also influences biomass estimates. Damaged specimens with punctured cuticles, or even lacking appendages and other body parts can lose indefinable amounts of body fluids during preservation, which introduces an immeasurable error into biomass estimation (Gaston et al. 1996). Especially, fragile taxa without protective exoskeletons, such as annelids, are more likely to suffer from such treatment than others. Loss of body fluid can only be partially compensated for when the entire preservative is included in DW and AFDW estimation. Dermott and Paterson (1973) found significantly higher DW estimates when the entire preservative solution was allowed to dry with the samples. Likewise, filling state of the animals alimentary tract can account for some extra weight. In chironomids, for examples, detritus and sand grains can amount for 4% of the measured DW (Landahl and Nagell 1978), a factor relevant especially when deposit feeding specimens are investigated. Moreover, gut content material such as sand grains affect estimates of DW and AFDW.
Considering all potential error sources mentioned, the effect of different preservatives seems to be negligible. In our opinion, biomass estimation is mainly affected by the sampling procedure itself, and by the sample handling rather than by the choice of the preservative. If sample treatment and biomass estimation are carried out with reasonable care, convenient conversion factors for preserved samples may be established for probably any species. When the same methods of preservation and sample handling are adopted as in the present paper, the conversion factors reported herein may be used. However, caution is advised for different species. Although our factors may be applicable for related species as well (same taxa and same size class), this has not been tested yet.
The use of ethanol as a preservative has, in addition to the lower health risk for operator/laboratory personal, also some additional advantages, e.g. if molecular determination of specimens using DNA samples is desired. Though, DNA amplification by polymerase chain reaction techniques (PCR) can be carried out using formalin fixed samples (e.g. Greer et al. 1991), this is difficult and error-prone (Quach et al. 2004). Besides, spills, which can occur in the heat of sampling, can be taken with much more calmness because the fixative is less hazardous to man and nature. However, we do not want to leave unmentioned that ethanol use is not without inconvenience. First, for the same amount of samples much more fixative solution is required using ethanol (96%) to obtain the final (70%) concentration, especially when the samples contain high portions of water. Using a formalin stock solution (e.g. 36%) only small amounts need to be added to samples to achieve the desired 4% final concentration. Second, the high vapor pressure of ethanol can make regular refilling of sample vials necessary (Black and Dodson 2003); and third, and probably most important, ethanol preserved specimen tend to be rather soft and mushy compared to formalin-fixed samples so that they are not usable for museum collections and care has to be taken not to harm fragile body parts or to puncture the cuticle during post fixative sample handling.
In this study, we have presented compelling evidence that the choice of the preservative has little effect on biomass loss. Both preservatives result in pronounced losses of biomass estimates, which however, is largely restricted to the first 3 weeks. Several other factors (e.g. sampling and sample handling) affect the precision of biomass estimation probably much more than the choice of the preservative. We conclude that the use of the toxic formalin solution should be abandoned, or at least limited to the few useful applications (e.g. museum collections and meiofauna fixation). The considerable advantage is clear: lower health risks for scientists and laboratory personnel. Furthermore, our biomass conversion factors provide other researchers with a technique to calculate the pre-fixative biomass in four estuarine macroinvertebrates from preserved specimens. These factors may even be applied to other related species within the same size category, but further research is welcome to test this hypothesis and to obtain values for other benthic species. Especially values from other common marine and estuarine invertebrates would be an excellent completion.
We gratefully acknowledge the good cooperation with the owners of the Bioconsult Schuchardt & Scholle GbR consulting company, Bremen, Dr. Bastian Schuchardt and Jörg Scholle. The company collected individuals, performed biomass estimations, and provided help with discussions on our account. We also thank our colleagues Dr. Andrea Warbeck (University of Hamburg) and Martina K. Sauer (Federal Institute of Hydrology) and two anonymous reviewers for their constructive reviews of the manuscript.
- AIHA (1989) Odor thresholds for chemicals with established occupational health standards. American Industrial Hygiene Association, Fairfax, VAGoogle Scholar
- Black AR, Dodson SI (2003) Ethanol: a better preservation technique for Daphnia. Limnol Oceanogr Methods 1:45–50Google Scholar
- Cavaletto J, Nalepa T, Fanslow D, Schlosser D (2003) Temporal variation of energy reserves in mayfly nymphs (Hexagenia spp.) from Lake St Clair and western Lake Erie. J Freshwat Biol 48:1726–1738View ArticleGoogle Scholar
- Crawley MJ (2002) Statistical computing. An introduction to data anlysis using S-Plus. Wiley, New YorkGoogle Scholar
- Dermott RM, Paterson CG (1973) Determing dry weight and percentage dry matter of chironomid larvae. Can J Zool 52:1243–1250Google Scholar
- Gaston GR, Bartlett JHW, Mcallister AP, Heard RW (1996) Biomass variations of estuarine macrobenthos preserved in ethanol and formalin. Estuaries 19:674–679Google Scholar
- Greer C, Peterson S, Kiviat N, Manos M (1991) PCR amplification from paraffin-embedded tissues. Effects of fixative and fixation time. Am J Clin Pathol 95:117–124PubMedGoogle Scholar
- Hamilton S, Kingston PF (1985) The effects of the preservatives, alcohol, formalin and propylene phenoxetol on the wet weights of some marine animals, Report to the ICES Benthos Working Group: 2Google Scholar
- HELCOM (1988) Guidelines for the Baltic Monitoring Programme for the Third Stage; Part D. Biological Determinands. Balt Sea Environ Proc 27D:164Google Scholar
- Howmiller RP (1972) Effects of preservatives on weights of some common macrobenthic invertebrates. Trans Am Fish Soc 101:743–746View ArticleGoogle Scholar
- ICES (1994) Report of the ICES/HELCOM workshop on quality assurance of benthic measurements in the Baltic Sea. ICES CM. 1994/E:10, Kiel, GermanyGoogle Scholar
- ICES (1996) Report of the ICES/HELCOM second workshop on quality assurance of benthic measurements in the Baltic Sea. ICES CM. 1996/E:2, Helsinki, FinlandGoogle Scholar
- Ihaka R, Gentleman R (1996) R: a language for data analysis and graphics. J Cmp Gr St 5:299–314Google Scholar
- Jonasson PM (1972) Ecology and production of the profundal benthos in relation to the phytoplankton in Lake Esrom Oikos. (Suppl14):148Google Scholar
- Landahl C-C, Nagell B (1978) Influence of the season and the preservation methods on wet- and dry weights of larvae of Chironomus plumosus L. Int Revue Ges Hydrobiol 63:405–410Google Scholar
- Leuwen SEW, Brock CM, Druten HAM (1985) Effects of preservation on dry-free and ash-free dry weight biomass of some common aquatic macro-invertebrates. Hydrobiology 127:151–159View ArticleGoogle Scholar
- Mason WT, Lewis PA, Weber CI (1983) An evaluation of benthic macroinverbrate biomass methology. Environ Monit Assess 3:29–44View ArticleGoogle Scholar
- OSPARCOM (1997) JAMP eutrophication monitoring guidlines: BenthosGoogle Scholar
- Quach N, Goodman MF, Shibata D (2004) In vitro mutation artifacts after formalin fixation and error prone translesion synthesis during PCR. BMC Clin Pathol 4: DOI 10.1186/1472-6890-4-1View ArticleGoogle Scholar
- Ruhmor H (1990) Soft bottom macrofauna: collection and treatment of samples. ICES. Tech Mar Environ Sci 8:18Google Scholar
- Ruhmor H, Brey T, Ankar S (1987) A compilation of biometric conversion factors for benthic invertebrates of the Baltic Sea. The Baltic Marine Biologists Publ 9:56Google Scholar
- Salonen K, Sarvala J (1985) Combination of freezing and aldehyde fixation, a superior preservation method for biomass determination of aquatic invertebrates. Arch Hydrobiol 103:217–230Google Scholar
- Sokal RR, Rohlf FJ (1998) Biometry: The principles and practice of statistics in biological research. W.H. Freeman and Company, New YorkGoogle Scholar
- Wiederholm T, Eriksson L (1977) Effects of alcohol-preservation on the weight of some benthic invertebrates. Zoon 5:29–31Google Scholar