Preserving fish specimens in museum jars quietly rewrites the very numbers scientists use to study them. A ten-year experiment on the large-lip killifish, Tlaloc labialis, a small freshwater species endemic to streams of southern Mexico and Guatemala, shows that standard preservation practices permanently distort body weight, length, and the statistical relationships built on them. Researchers followed the same fifty fish from the moment of capture through formalin fixation and a full decade in ethanol, and the results carry a blunt warning for anyone who treats museum collections as a faithful record of what fish once weighed in life.
The specimens came from tributaries of the Grijalva River basin in Chiapas, Mexico, where they were collected in 2015 with a seine net measuring three by two meters with half-centimeter mesh. After capture, the fish were euthanized with an overdose of MS-222, consistent with guidelines for the use of fishes in research published by the American Fisheries Society, and field activities were authorized by the Mexican Ministry of Environment and Natural Resources under permit SGPA/DGVS/02208/13. In the laboratory, excess moisture was blotted away with absorbent paper, individuals were carefully straightened when needed, and each fish was measured and weighed while fresh. The animals were then fixed in ten percent formalin for seventy-two hours at room temperature before being transferred to seventy percent ethanol for final preservation and accessioning in the Fish Collection of the Zoology Museum of the University of Sciences and Arts of Chiapas, under catalog numbers MZ-UNICACH 07813 and 07814, with twenty-five specimens each.
Because the study used a longitudinal, repeated-measures design, the exact same fifty individuals were monitored at every stage. Standard length and total weight were recorded when the fish were freshly collected, seventy-two hours after formalin fixation, one month after transfer to ethanol, and again after one, two, five, seven, nine, and ten years in ethanol. Before each measurement, excess moisture was removed by blotting. Overall standard length ranged from 3.14 to 8.20 centimeters, and based on the size at first sexual maturity of 4.10 centimeters reported in earlier work, thirteen fish were classified as mature and thirty-seven as immature. This repeated-measures structure matters: rather than comparing different jars of fish at different ages, the researchers tracked individual bodies through time, which allowed them to attribute each change directly to the chemistry of preservation.
The headline finding is a steep and largely one-way loss of mass. Freshly caught specimens lost an average of 5.78 percent of their weight, with a standard deviation of 3.01, within seventy-two hours of formalin fixation. The decline then accelerated dramatically during the first month in ethanol, reaching minus 29.94 percent, with a standard deviation of 5.35, before stabilizing at minus 33.92 percent after one year. From that point on, the weight loss essentially stopped, with no further significant losses observed through the end of the ten-year study period. Statistically, the differences among preservation stages were overwhelming: a Friedman test produced a chi-squared value of 311.85 with eight degrees of freedom and a P-value below 0.001, and post hoc Wilcoxon signed-rank tests with Bonferroni correction confirmed that the most pronounced changes occurred during the initial stages of fixation and preservation. The same overall pattern appeared for length, though on a much smaller scale.
That smaller scale is itself informative. During initial fixation, specimens shrank linearly by 1.19 percent in standard length, a figure that increased about 2.1-fold after the first month in alcohol to 2.53 percent. Longitudinal shrinkage then advanced gradually, reaching 3.04 percent at one year of preservation, after which length stabilized with no further significant shrinkage through year ten. In other words, the fish lost roughly a third of their weight but only about three percent of their length. This asymmetry is the engine behind the study’s most consequential result, because length–weight relationships are built on exactly this kind of paired data. When mass falls far faster than length, the slope of that relationship must steepen, and the steepening is an artifact of the jar, not of the fish’s biology.
The length–weight relationship, expressed in the allometric equation W = aL^b, where W is weight in grams and L is standard length in centimeters, is a fundamental tool in fisheries and natural resource management, providing essential data for biomass estimation, growth trajectories, assessment of fish stock condition, and conservation studies. It can be used to estimate fish biomass at a given length and to develop behavioral or predictive models at both population and community levels. In this population, the exponent b ranged from 3.26, with a 95 percent confidence interval of 3.184 to 3.336, in fresh specimens to 3.574, with an interval of 3.469 to 3.678, at two years, settling at 3.457, with an interval of 3.344 to 3.571, after ten years. The b-value was above 3 at every stage, including in fresh specimens, indicating positive allometric growth in the living population. The increase was highly pronounced from fresh specimens to those after one month in ethanol, and the ten-year monitoring confirmed that while the initial shift was rapid, a final adjustment occurred between the first month and the first year of ethanol preservation, with no further significant changes during subsequent annual intervals. Confidence intervals for b in fresh and fixed specimens did not overlap with those of the other periods, except for fresh specimens versus one year in ethanol.
Fulton’s condition factor, K, calculated as the weight divided by the cube of length multiplied by one hundred, offers a proxy of fish health and plumpness, resting on the assumption that fish weighing more at a given length are in better condition than those weighing less at the same length. Here, K fell from 2.08 in freshly caught fish to 2.03 in formalin-fixed specimens after seventy-two hours, and the decline continued to 1.51 after one year in ethanol, after which values remained relatively constant. Mean K-values differed significantly among the nine time periods evaluated, with a Friedman chi-squared of 231.7 and a P-value below 0.001. The authors interpret this marked decline as a direct consequence of mass loss and a preservation artifact, not a true indication of poor biological health or of environmental stressors affecting the sampled population. That distinction is critical for biomonitoring, where condition factors are routinely read as signals of environmental quality.
The distortion was not uniform across the sample. Although both maturity groups followed the same overall temporal trend of weight loss, a distinct and constant gap emerged: smaller, immature individuals proved more susceptible to preservation-induced shrinkage, stabilizing at roughly 35 percent weight loss, while mature specimens maintained a significantly lower loss of around 26 percent across the ten-year monitoring period. A linear mixed model, with maturity stage, measurement, and their interaction as fixed effects and individual fish identity as a random effect, confirmed these differences were statistically significant, with F(1, 48) = 44.74 and P < 0.001 for the maturity effect and F(7, 336) = 7.74 and P < 0.001 for the maturity-by-measurement interaction. This size-dependent desiccation artificially steepens the slope of the length–weight relationship, a mechanism the authors note is likely accelerated by complementary processes such as differential longitudinal shrinkage among size classes and the combined effects of tissue dehydration and overall body contraction. The chemistry is straightforward: formalin acts as a fixative primarily through protein cross-linking, while ethanol serves as a potent dehydrating agent that replaces biological water within the tissues, driving severe mass loss.
The practical payoff comes in the form of correction equations. Based on the stabilized overall mass loss of minus 35.5 percent for immature specimens and minus 26.8 percent for mature specimens observed after one year, the researchers propose two formulas to estimate fresh weight from preserved weight for fish kept one year or more in alcohol: for mature specimens, fresh weight equals preserved weight divided by 0.732, and for immature specimens, fresh weight equals preserved weight divided by 0.645. These equations apply only to specimens preserved for at least one year in seventy percent ethanol following seventy-two hours of formalin fixation. The authors caution that biomass loss is size-dependent, so researchers should bear the maturity-specific pattern in mind, but the factors allow approximate back-calculation of original biomass from historical museum series. The findings also push back against the suggestion that preservation-induced biases are negligible compared with natural seasonal oscillations in growth; here, desiccation demonstrably altered both the condition factor and the length–weight exponent. At the same time, the authors agree that museum collections provide an invaluable framework for deriving key ecological parameters without new field sampling, and they note that after the initial losses during the first year, the biometric indices of these specimens remained stable for at least a decade, a pattern that may also hold for closely related species. For a species with prior baselines showing that chemical preservation artificially increases the allometric coefficient, this study extends that warning across a full ten years and turns it into a usable quantitative correction.
Subject of Research: Long-term effects of formalin and ethanol preservation on biometric indices of the large-lip killifish
Article Title: Long-term trends of biometric indices in the large-lip killifish, Tlaloc labialis ( Actinopterygii , Cyprinodontiformes , Profundulidae )
Article References: Anzueto-Calvo, M. D. J., Cruz-Maza, B., Domínguez-Cisneros, S. E., & Velázquez-Velázquez, E. (2026). Long-term trends of biometric indices in the large-lip killifish, Tlaloc labialis (Actinopterygii, Cyprinodontiformes, Profundulidae). Acta Ichthyologica et Piscatoria, 56, 475–481. https://doi.org/10.3897/aiep.56.199022
Image Credits: AI Generated
Keywords: Tlaloc labialis, large-lip killifish, length–weight relationship, Fulton's condition factor, museum specimens, ethanol preservation, formalin fixation, biomonitoring, fisheries management, Grijalva River basin, specimen shrinkage, correction factors
News Source: Violet Maxwell. (November 5, 2026). Museum Pickling Warps Fish Body Measurements for a Decade, Study Finds. Scienmag.
