Length-weight relationships of 24 stream-dwelling fish species from the Atlantic Forest in Rio de Janeiro, Brazil

Abstract The length-weight relationship parameter is important for obtaining fish weight and biomass data with relevant implications about species role on ecosystem functioning. Here we report the length-weight relationship (LWR) for 24 fish species from three streams located in the Atlantic Forest in the Rio de Janeiro State, Brazil. Fish were collected with electrofishing and standard length (cm) and wet weight (g) were measured to obtain the a and b parameters of the Log (W) = Log (a) + b Log (SL) equation. Length-weight relationships for seven out of 24 species (Hypostomus punctatus, Deuterodon taeniatus, Deuterodon hastatus, Deutorodon janeiroensis, Characidium vidali, Characidium interruptum and Rineloricaria zawadiskii) are reported for the first time. The length-weight relationships reported here contribute to the database that can support fish diversity conservation, fisheries management plans and studies on fish biology.


Introduction
Length-weight relationships (LWR) are very important for ecological, ichthyological and fisheries studies as it allows one to calculate weight by measuring length and to estimate the condition factor, providing important information on fish physiology (Le Cren 1951, Froese 2006).Data about LWR can also elucidate different aspects about fish health, individual growth and, once the fish weight is known, it is possible to estimate the biomass of an entire population or community and thus investigate ecosystem processes and how fish species can affect them (Froese 2006, Giarrizzo et al. 2006, Joyeux et al. 2009, Camara et al. 2011, Zandonà et al. 2021, Lima et al. 2022).
Here, we report the length-weight relationships for 24 streamdwelling fish species from three Atlantic Forest streams located in the state of Rio de Janeiro, Brazil.These fish species are relatively common in the Atlantic Forest streams, but information on length-weight relationships are missing for seven species (Hypostomus punctatus, Deuterodon taeniatus, Deuterodon hastatus, Deuterodon janeiroensis, Characidium vidali, Characidium interruptum and Rineloricaria zawadiskii) in FishBase (Froese & Pauly 2023) or other published literature.As a matter of fact, Rineloricaria zawadiskii was described for the first time in 2022, therefore this is the first data on LWR for this species (Silva et al. 2022).In this sense, our results can be of great importance for all researchers working in this biome.

Material and Methods
Fish were collected in three Atlantic Forest streams in Rio de Janeiro, Brazil: Rio Guapiaçu (22°26′08.2′′Sand 42°45′31.8′′W),Rio Ubatiba (22°52′15.9′′Sand 42°44′14.1)and Rio Mato Grosso (22°52'24.7"Sand 42°39'06.4"W)(Figure 1).Rio Guapiaçu is within the Guapiaçu-Macacu basin, which provides water to 2.5 million people among five cities in the surrounding area.Rio Guapiaçu is a fourthorder stream with a maximum width of 13m and a maximum depth of 1.5m (Manna et al. 2019).The riparian area is densely forested and the substrate composed of a mixture of bedrock, sand and leaf litter distributed across runs and pools (Manna et al. 2017).Rio Ubatiba is a low-elevation second-order stream in the Serra do Mar mountain   Fish sampling was conducted in Rio Guapiaçu during the dry season in June, July, and October 2012; in Rio Ubatiba during three periods: (i) from July 1994 to July 1995, (ii) July 2015 to January 2016, and (iii) February 2019; and in Rio Mato Grosso from March 2016 to January 2017.In all sites, fish were collected in a 200-m long reach, delimited by two seine nets (mesh size 5 mm).We used a backpack portable electrofishing device (Mazzoni et al. 2000) to catch fish using successive removals (Zippin 1958).The fish were euthanized with a solution of 0.4 ml of eugenol, 3.6 ml of methanol and 1L of distilled water, and subsequently fixed in formalin 10%.After seven days, all fish specimens were preserved in 70% ethanol.

Characidae
Fish species identification was confirmed by specialists at the National Museum of the Federal University of Rio de Janeiro (MN-UFRJ).After fixation, fish were measured with a caliper for standard length (to the nearest 0.01 cm) and weighted (wet weight, nearest 0.01g).The length-weight relationships were obtained using the equation Log (W) = Log (a) + b Log (SL) where W is the fish wet weight (g), and SL is the standard length (cm).The coefficient of determination (R 2 ) and the 95% confidence interval of the a and b parameters of the equation were calculated.Outliers were excluded by visual detection when plotting the Log W -Log SL relationship.

Results
In this study, we sampled a total of 1081 specimens belonging to 24 fish species and 11 families to obtain their length (cm) and weight (g) measurements.Length-weight relationships (LWR) from each species and study site, sample sizes (N), ranges for standard length and weight, the means and 95% Confidence Interval (95% CI) for the equation parameters a and b, and the coefficient of determination (R 2 ) are presented in Table 1.The coefficient of determination (R 2 ) ranged from 0.720 to 0.994, a values ranged from 0.0008 to 0.039 and b values from 2.55 to 3.60. Continue...

Discussion
In this study, we present length-weight relationships (LWR) of 24 fish species, of which seven have no previous LWR documented in FishBase (Froese & Pauly 2023).The coefficient of allometry (b) ranged from 2.55 (Phalloceros harpagos) to 3.74 (Rhamdia quelen ), with the majority of the species presenting values within the expected range of 2.5-3.5 (Froese 2006).
In contrast to our results, previous studies described higher values of standard length, R 2 and b coefficients for P. harpagos, revealing that differences in local conditions, such as habitat, seasonal variation and stream area, can explain differences in length-weight relationships (Campos et al. 2020).Other factors, such as predation and intra and interspecific competition can affect reproduction, condition factor, and growth rate, thus influencing maximum length, intraspecific weight variation, and a and b coefficients, as observed here.For instance, P. harpagos fecundity and size is affected by predation (Gorini-Pacheco et al. 2018).Also, being a matrotrophic species (the mother provides nutrients to the developing embryos throughout pregnancy), pregnancy stage could affect the female weight (Zandonà et al. 2021) and thus be a source of variation in the LWR.
Moreover, our study observed a low R 2 for P. maculicauda in a stream impacted by pasture (R 2 = 0.720).In contrast, Campos et al (2020) found a R 2 = 0.956 for the same species in well preserved streams.This difference in R 2 values could indicate the influence of environmental conditions in LWR.Therefore, it is possible that the lower R 2 in impacted streams could be a consequence of differences in resource availability which is resulting in higher individual variability in fish weight.Other sources of intraspecific variation in fish weight that could be causing lower R 2 values are differences in reproductive status, condition factor, or seasonality (Campos et al. 2020, Lima et al. 2022).As for S. marmoratus, it is worth mentioning that its low sample size (Rio Guapiaçu: n = 5; Rio Ubatiba: n = 7) is because this is a rare species.Thus, this estimate must be treated with caution.
Anthropogenic threats to fish biodiversity from Atlantic Forest coastal streams are already known, caused especially by biological invasions and environmental alterations (Bezerra et al. 2019).Therefore, it is very important to gather information and increase our understanding of fish biology to trace effective conservation strategies and preserve the biodiversity of this biome.

Figure 1 .
Figure 1.Map of the study area in the state of Rio de Janeiro, southeastern Brazil.Black dots = sampling sites: (a) Guapiaçu stream, municipality of Guapimirim, (b) Ubatiba and Mato Grosso streams, municipality of Maricá and Saquarema, respectively.

Table 1 .
Length-weight relationships (LWR) for 24 fish species from Atlantic Forest streams in Brazil.
It flows through meadows deforested for agricultu, Manna et al. 2020attle ranching, but patches of secondary Atlantic Forest are still common on the top of the surrounding hills.The average width and depth of our sampling sites was 1.2 m and 0.4 m, respectively.Rio Mato Grosso is a third-order stream that flows for about 11 km before discharging into the Saquarema lagoon, a large brackishwater lagoon formed by the accumulation of sand dunes offshore.Average width and depth of our sampling sites was 3.6 m and 27.4 cm, respectively (this sampling site is described in more detail inMazzoni & Lobón-Cervià  2000, Rezende et al. 2013, Manna et al. 2020).
. On the other hand, for P. harpagos, S. barbatus, H. malabaricus, M. microlepis, Jenynsia multidentata and P. maculicauda, b values are lower than Fishbase data, while for A. multispinis, Poecilia vivipara and G. brasiliensis, b values are higher.This variability in b values is also present in previous studies and can a combination of biological factors, such as the length range of individuals sampled, their maturity stage, stomach fullness, diet, the presence of disease or parasite, and environmental conditions, such as temperature, pollution or fishing pressure