Marine and Fishery sciences 40 (1): xxx-xxx (2027)

https://doi.org/10.47193/mas4012027010103

Marine and

Fishery Sciences

MAFIS

*Correspondence:

jdnunez@mdp.edu.ar

Received: 5 August 2026

Accepted: 1 October 2026

ISSN 2683-7595 (print)

ISSN 2683-7951 (online)

https://ojs.inidep.edu.ar

Journal of the Instituto Nacional de

Investigación y Desarrollo Pesquero

(INIDEP)

This work is licensed under a Creative

Commons Attribution-

NonCommercial-ShareAlike 4.0

International License

ORIGINAL RESEARCH

Articial light at night and apparent predation on juvenile Neohelice

granulata during two deployment periods at Mar Chiquita, Argentina

Jesús d. nuñez1, *, ruben a. robles1, Macarena Pérez García1, caMila b. Merodio2,

M. GuadaluPe Vázquez1 and ToMás a. luPPi1

1Instituto de Investigaciones Marinas y Costeras (IIMyC), Facultad de Ciencias Exactas y Naturales (FCEyN),Universidad Nacional de Mar

del Plata (UNMdP), Consejo Nacional de Investigaciones Cientícas y Técnicas (CONICET), Juan B. Justo 2550, B7608FBY - Mar del

Plata, Argentina. 2Instituto de Biología de Organismos Marinos (IBIOMAR), Nacional de Investigaciones Cientícas y Técnicas (CONICET),

Abuelas de Plaza de Mayo 2915 - U9120ACD, Puerto Madryn, Argentina. ORCID Jesús D. Nuñez https://orcid.org/0000-0001-9593-2449,

Ruben A. Robles https://orcid.org/0009-0003-5496-7604, Macarena Pérez García https://orcid.org/0000-0002-0174-3004,

Camila B. Merodio https://orcid.org/0009-0001-9858-7035, M. Guadalupe Vázquez https://orcid.org/0000-0003-0572-3796,

Tomás A. Luppi https://orcid.org/0000-0002-4959-5987

ABSTRACT. Over the past decade, articial light at night (ALAN) has become one of the most

spatially extensive anthropogenic disturbances along coastlines worldwide, yet whether its capacity

to alter predation is a xed property of illumination intensity or depends on the ecological context in

which it occurs has received little attention. Here we evaluate this question using replicated tethering

experiments at the microtidal lagoon of Mar Chiquita (south-western Atlantic coast of Argentina) in

two deployment periods (spring and summer), quantifying apparent predation on juvenile Neohelice

granulata under ALAN (cold-white LED, ≈ 38 lux) and natural-darkness controls. A Generalised

Linear Mixed Model with experimental night as a random intercept revealed a signicant interaction

between light condition and season. In spring, ALAN signicantly increased the probability of juve-

nile predation relative to natural darkness, whereas no detectable effect was found in summer. Placed

alongside the published effect at a macrotidal, structurally simpler site, ALAN sensitivity follows

an ordered gradient consistent with an amplifying-effect hypothesis, whereby ALAN would amplify

predation in proportion to the underlying susceptibility state of predators and prey (adult nutritional

condition, juvenile developmental vulnerability and habitat structural complexity) rather than being a

xed property of light intensity alone, although adult nutritional condition was not measured directly

and is inferred from published seasonal values for this population. These results indicate that the

same habitat may shift between high- and low-sensitivity states across seasons, that single-season

assessments may systematically underestimate cumulative predation exposure, and that targeted light

management during juvenile settlement is, in principle, a more efcient conservation approach than

year-round restrictions, although we did not evaluate its costs or feasibility.

Key words: Predation risk, light pollution, coastal urbanization, predator-prey interactions.

Luz articial nocturna y depredación aparente sobre juveniles de Neohelice granulata durante

dos períodos de despliegue en Mar Chiquita, Argentina

RESUMEN. En la última década, la luz articial nocturna (ALAN, por sus siglas en inglés)

se ha convertido en una de las perturbaciones antropogénicas de mayor extensión espacial en las

costas de todo el mundo; sin embargo, se ha prestado poca atención a si su capacidad para alterar la

depredación es una propiedad ja de la intensidad lumínica o si depende del contexto ecológico en

el que se produce. En este estudio evaluamos dicha cuestión mediante experimentos replicados de

sujeción de presas (tethering) en la laguna micromareal de Mar Chiquita (costa atlántica del suroeste

Marine and Fishery sciences 40 (1): xxx-xxx (2027)

INTRODUCTION

The Earth’s night is getting brighter at a rate

of 6% per year because of the global expansion

of articial light at night (ALAN) (Hölker et al.

2010). Over the past two decades, ALAN has

emerged as one of the most spatially extensive an-

thropogenic disturbances, with notorious impacts

on populations and communities in both terrestrial

and marine ecosystems (Davies et al. 2014; Gas-

ton et al. 2015; Bennie et al. 2018; Sanders et al.

2018). Coastal and intertidal zones are among the

most affected, given the continuing urbanization of

coastlines worldwide (Perkin et al. 2011; Garratt et

al. 2019; Zapata et al. 2019). ALAN’s direct mod-

ication of natural lightscapes may have critical

consequences for species interactions by disrupting

the biological rhythms, locomotion patterns, and

refuge-seeking behaviors of benthic vertebrates

(Bolton et al. 2017) and invertebrates (Underwood

et al. 2017; Nuñez et al. 2021b) and modifying the

environmental contexts under which predator-prey

encounters occur (see Davies et al. 2013; Gaston et

al. 2013; Bolton et al. 2017). In particular, articial

illumination can extend the foraging window of

visually oriented predators, increase prey detection

efciency, and erode the diel temporal refuge that

prey species rely on for safety (e.g. Cerri 1983;

Batty et al. 1990; Bolton et al. 2017; Sanders et al.

2018; Czarnecka et al. 2019; McMahon et al. 2022).

Nevertheless, the outcome is not uniform: under

broad-spectrum white light, a shore crab attacked

fewer snails overall and the pattern of color-based

prey selection was reversed (Moyse et al. 2023),

which indicates that the consequences for predation

depend on the species and the ecological context.

Although the effects of ALAN on predation

have recently begun to be documented in several

taxa (see Hampton 2004; Underwood et al. 2017;

Sanders et al. 2018), whether the magnitude of

ALAN-driven predation amplication is a xed

function of light intensity or depends on the ecolog-

ical context in which illumination occurs remains

poorly understood. If ALAN sensitivity varies pre-

dictably with measurable ecological attributes such

as habitat complexity, nutritional state of predators,

or developmental stage of prey, then risk assess-

ment becomes tractable and management interven-

tions could be prioritized accordingly. If it does not,

any single-site, single-season measurement may

misrepresent the broader impact.

In size-structured populations of benthic inver-

tebrates, where adults cannibalize recently settled

juveniles (Polis 1981; Elgar and Crespi 1992), the

de Argentina) durante dos periodos de muestreo (primavera y verano), cuanticando la depredación aparente sobre juveniles de Neohe-

lice granulata bajo condiciones de ALAN (LED de luz blanca fría, ≈ 38 lux) y controles de oscuridad natural. Un modelo lineal mixto

generalizado, con la noche experimental como intercepto aleatorio, reveló una interacción signicativa entre la condición lumínica y la

estación. En primavera, ALAN aumentó signicativamente la probabilidad de depredación de juveniles en comparación con la oscuridad

natural, mientras que en verano no se observó ningún efecto detectable. Al comparar estos resultados con los publicados para un sitio

macromareal de estructura más simple, se observa que la sensibilidad a la ALAN sigue un gradiente ordenado coherente con la hipótesis

de un efecto amplicador: ALAN amplicaría la depredación en proporción al estado de susceptibilidad subyacente de depredadores

y presas (condición nutricional de los adultos, vulnerabilidad del desarrollo de los juveniles y complejidad estructural del hábitat), en

lugar de ser una propiedad ja derivada únicamente de la intensidad de la luz (si bien la condición nutricional de los adultos no se midió

directamente, sino que se inrió a partir de valores estacionales publicados para esta población). Estos resultados indican que un mismo

hábitat puede alternar entre estados de alta y baja sensibilidad según la estación, que las evaluaciones realizadas en una sola estación

pueden subestimar sistemáticamente la exposición acumulada a la depredación y que la gestión especíca de la iluminación durante el

asentamiento de los juveniles constituye, en principio, un enfoque de conservación más eciente que las restricciones aplicadas durante

todo el año, aunque no evaluamos los costos ni la viabilidad de dicha medida.

Palabras clave: Riesgo de depredación, contaminación lumínica, urbanización costera, interacciones depredador-presa.

Nuñez et al.: alaN modulates predatioN iN Neohelice graNulata 3

conditions generating ALAN sensitivity may be

particularly dynamic. Cannibalism intensity de-

pends on the interaction among adult motivation

(nutritional state, reproductive energetics), juve-

nile vulnerability (developmental stage, refuge

establishment), and habitat structure (refuge den-

sity, visual complexity of the substrate; Luppi et al.

2001; Claessen et al. 2004; Moksnes 2004; Hines

et al. 2026). Each of these parameters changes both

seasonally and geographically, so the same popu-

lation may shift between high- and low-sensitivity

states as the year progresses. A 37-year tethering

experiment on the blue crab Callinectes sapidus

has examined cannibalism, predation and refuges

of juveniles (Hines et al. 2026), and tethering can

serve as a comparative index of predation when

applied consistently across treatments (Aronson

and Heck 1995; Baker and Waltham 2020).

The burrowing crab Neohelice granulata (Dana,

1851) provides an ideal model system to evaluate

context-dependence in ALAN-predation interac-

tions. This species is the dominant ecosystem en-

gineer of south-western Atlantic salt marshes, dis-

tributed along more than 4000 km of Argentine and

Brazilian coast (Spivak et al. 1994), and modulates

sediment biogeochemistry, organic-matter cycling,

and benthic community structure through its exten-

sive burrowing activity (Botto and Iribarne 2000;

Alberti et al. 2015; Martinetto et al. 2016). Adults

prey heavily on recently settled conspecics, with

cannibalism strongly mediated by refuge availabil-

ity, predator hunger, and juvenile developmental

stage (Luppi et al. 2001; Méndez Casariego et al.

2009, 2011). Nuñez et al. (2021a) provided the rst

eld evidence that ALAN substantially amplies

juvenile mortality in N. granulata at the macrot-

idal, structurally simple site of San Antonio Oeste

(SAO), where the combination of open sediment

ats and chronically nutritionally stressed adult

populations (Bas et al. 2019) seems to generate

extreme ALAN sensitivity. A related study fur-

ther documented that N. granulata exhibits strong

behavioural plasticity in its ALAN responses de-

pending on ecological context (Nuñez et al. 2021b).

Whether such sensitivity extends to the structurally

complex, high-quality lagoon habitat of Mar Chiq-

uita (MCH), and whether it varies across seasons,

had not been tested.

Although the general prediction that ALAN

enhances predation by extending the foraging

window of visual predators is well supported (e.g.

Becker et al. 2013; Bolton et al. 2017; Sanders et

al. 2018; Czarnecka et al. 2019), the magnitude of

this effect may not be independent of the predatory

context of the system. In this sense, we identify

four non-exclusive, heuristic hypotheses (not mod-

els formally discriminated by the present design)

about how ALAN effects on predation interact with

the underlying susceptibility state of the system,

understood as the combined outcome of adult nu-

tritional condition, juvenile developmental vulner-

ability, and habitat structural complexity, for which

within-site baseline (no-ALAN) predation provides

an empirical proxy (Figure 1): (HA) Additive ef-

fect. ALAN adds a constant increment to predation

regardless of the susceptibility state of the system,

implying no interaction between light pollution

and the predatory state of the system (see Gaston

et al. 2013; Bolton et al. 2017). (HB) Amplifying

effect. The ALAN effect scales with susceptibility:

systems already characterized by high predatory

pressure would experience the greatest absolute

increase under ALAN, for instance if nutritionally

stressed or behaviorally active predators exploit the

extended foraging window more intensively (e.g.

Becker et al. 2013; Sanders et al. 2018; Czarnecka

et al. 2019), as has been suggested for protein-lim-

ited populations of N. granulata at SAO (Bas et al.

2019). (HC) Compensatory effect. ALAN effects

are largest in low-susceptibility systems, where

prey detection under natural darkness is most con-

strained. This implies that habitat complexity and

resource availability may buffer ALAN’s inuence

(Luppi et al. 2001; Méndez Casariego et al. 2009).

(HD) non-linear (threshold) effect. The ALAN

effect may peak at an intermediate susceptibility

state: too few motivated predators in low-suscep-

tibility systems, while high-susceptibility systems

Marine and Fishery sciences 40 (1): xxx-xxx (2027)

could already be behaviorally saturated (see Gas-

ton et al. 2013). These hypotheses are not mutually

exclusive and may operate at different spatial or

temporal scales. Our design combines a within-site

seasonal contrast at MCH, where susceptibility was

expected to differ between a higher-risk spring pe-

riod and a lower-risk summer period, with a con-

textual comparison against the published, chroni-

cally high-susceptibility population at San Antonio

Oeste (SAO; a macrotidal site with limited habitat

complexity and protein-limited adults; Bas et al.

2019). This design provides an empirical basis to

evaluate which hypothesis best describes ALAN’s

inuence on predation in N. granulata, while rec-

ognizing that the SAO comparison draws on an

independent study and is treated as a qualitative

Figure 1. Conceptual hypotheses for how ALAN may modify predation as a function of the susceptibility state of the system. The

x-axis (‘baseline predation’) indexes this susceptibility state via no-ALAN predation intensity; within a single site and

protocol this is a direct empirical measure, whereas across independent studies (as for the San Antonio Oeste comparison

in this manuscript) it should be interpreted qualitatively rather than as a strictly commensurable metric (see Caveats and

limitations). A) Additive effect (HA): ALAN adds a constant increment to predation regardless of susceptibility, so the

ALAN and no-ALAN lines remain parallel. B) Amplifying effect (HB): the ALAN effect diverges from baseline as sus-

ceptibility increases, so contexts already characterised by high predation show the largest absolute increase under ALAN.

C) Compensatory effect (HC): the ALAN effect converges with baseline as susceptibility increases, so low-susceptibility

contexts show the largest relative increase under ALAN. D) Non-linear (threshold) effect (HD): the ALAN effect peaks at

an intermediate susceptibility state. In all panels, line colour indicates ALAN intensity (low, medium, high) and the black

dashed line marks the 1:1 reference (no ALAN effect).

Constant

effect

Baseline predation (ALAN)

Pre dation with ALAN

Intensity

High ALAN

Medium ALAN

LowALAN

No ALAN (reference)

Hypothesis A

Diverge

with baseline

Baseline predation (ALAN)

Pre dation with ALAN

Hypothesis B

Converge

with baseline

Baseline predation (ALAN)

Pre da tion with ALAN

Hypothesis C

Peak at

intermediate

Baseline predation (ALAN)

Pre dation with ALAN

Hypothesis D

Nuñez et al.: alaN modulates predatioN iN Neohelice graNulata 5

contextual anchor rather than a third quantitative

point on a common baseline-predation axis. Here

we report novel tethering data from two deploy-

ment periods at MCH: austral spring (September

2021), when we expected adult energy reserves

to be low and juveniles to be highly vulnerable,

and summer (February 2022), when we expect-

ed adults to be nutritionally replete and juveniles

to have had time to establish burrow refuges. We

predicted that ALAN would signicantly amplify

predation in spring (high susceptibility) but not in

summer (low susceptibility), and that the spring

MCH response would fall between the absent sum-

mer MCH effect and the extreme, chronic SAO

sensitivity, consistent with the amplifying hypoth-

esis (HB). We emphasize that baseline (no-ALAN)

predation is used here strictly as an empirical proxy

for the system’s overall susceptibility state, not as

a direct measurement of its underlying mechanis-

tic components; the candidate mechanisms linking

adult nutritional condition and juvenile vulnerabil-

ity to this pattern are developed in the Discussion

as post-hoc, non-measured interpretations, so as

to keep the classication of contexts (via baseline

predation) and the mechanistic explanation of the

observed pattern logically separate.

MATERIALS AND METHODS

Study system and sites

Neohelice granulata is a semiterrestrial burrow-

ing crab (adults 30-45 mm carapace width, CW)

that dominates intertidal sediments of south-west-

ern Atlantic salt marshes along more than 4,000 km

of Argentine and Brazilian coast. The juveniles

used in our experiments measured 12-14 mm

CW. Primary experiments were conducted at the

Mar Chiquita coastal lagoon (MCH; 37° 45' S, 57°

26' W; Figure 2), a Ramsar-listed microtidal sys-

tem (tidal range 0.3-0.8 m) on the Buenos Aires

Province coast of Argentina. Mar Chiquita is char-

acterized by organic-rich muddy sediments and a

dense, persistent burrow network of N. granulata

that provides structural microrefugia for juveniles.

Adults from Mar Chiquita maintain higher somatic

energy reserves than conspecics at structurally

simpler sites (Bas et al. 2019). Paired exclusion/

inclusion cage experiments at the structurally sim-

pler SAO site identied adult conspecics as the

dominant source of juvenile mortality there (Nuñez

et al. 2021a); no equivalent exclusion experiment

has been conducted at MCH, so predator identity

at this lagoon was inferred rather than directly es-

tablished (see Caveats and limitations). Published

tethering data from San Antonio Oeste (SAO; 40°

47' S, 65° 02' W; Figure 2), a macrotidal bay (tidal

range 5-7 m) characterized by exposed sand-cobble

intertidal ats with sparse burrow networks and

chronically protein-limited N. granulata popula-

tions (Bas et al. 2019; Nuñez et al. 2021a), were

used as a contextual reference only.

ALAN treatment

The ALAN was simulated using a cold-white

LED lamp (30 W, color temperature ≈ 6,000 K)

mounted on a 3 m PVC mast powered by a solar

panel, programmed to operate automatically from

dusk to dawn. At each site, multiple independently

located ALAN/control station pairs (‘blocks’) were

deployed simultaneously, each block consisting of

one ALAN station and one natural-darkness con-

trol station 30 m apart (Figure 2): four blocks in

summer and three in spring, each sampled across

ve and four experimental nights per season re-

spectively. Spring sampling was originally sched-

uled for ve nights; one night was excluded after

two of the three lighting units were stolen and the

remaining unit was damaged in the same incident,

leaving no usable data for that night. Control sta-

tions maintained < 0.1 lux and ALAN stations ≈ 38

lux at ground level throughout each night, as veri-

ed with a calibrated luxmeter (Mastech MS6610).

This intensity was chosen to match that used in

the companion San Antonio Oeste study (Nuñez

Marine and Fishery sciences 40 (1): xxx-xxx (2027)

et al. 2021a), enabling direct cross-site compari-

son, and corresponds to the mean level previously

measured under coastal lighting installations in the

study region. All experiments were conducted dur-

ing new-moon phases (± 3 d) to minimize ambient

lunar illumination.

Tethering experiments

A tethering experiment was conducted at MCH

in two deployment periods: austral spring 2021

(September; 4 usable experimental nights; 9 ALAN

and 9 control replicates; N = 88 crabs) and sum-

Figure 2. Study area and experimental design. Locator map of Argentina (left) showing the two study sites: Mar Chiquita (MCH;

red circle, this study) and San Antonio Oeste (SAO; green circle, published reference site, Nuñez et al. 2021a), each linked

by an arrow to a detailed inset map (right) giving latitude/longitude grid lines and a north arrow for that site. Top-right

panel: night-time photograph illustrating the ALAN experimental set-up, showing an unlit reference point and illuminated

(ALAN) stations, with the inter-station distance (30 m, matching the block spacing described in ALAN treatment) marked.

50 °S

45 °S

40 °S

35 °S

30 °S

25 °S

70 °W65°W60°W55°W

37.80

37.78

37.76

37.74

37.72

37.70

MAR CHIQUITA

40.74

40.73

40.72

40.71

SANANTONIO OESTE

ARGENTINA

Nuñez et al.: alaN modulates predatioN iN Neohelice graNulata 7

mer 2022 (February; 5 experimental nights; 23

ALAN and 23 control replicates; N = 223 crabs).

Juvenile N. granulata (CW 13 ± 1 mm; intermolt

individuals) were collected from unlit intertidal ar-

eas immediately before deployment to minimize

handling stress. Each juvenile was tethered with

25 cm transparent nylon monolament (0.20 mm

diameter) attached to the posterior carapace with

cyanoacrylate adhesive, which is biologically in-

ert once cured and is standard practice in crusta-

cean tethering studies (Baker and Waltham 2020),

and xed individually to its own sediment stake

(a 5 mm-diameter, 35 cm-tall plastic stake driven

into the sediment; one tethered juvenile per stake),

permitting free movement within a 25-cm radius

including partial burial. Three to ve stakes per

treatment and block were deployed each night, and

the outcomes of these stakes were pooled into a

single replicate (Table 1); crabs were deployed at

dusk and retrieved at dawn after one tidal cycle

(≈ 12 h). At retrieval, each individual was classi-

ed as: (i) alive and uninjured; (ii) predated (dead

with carapace damage characteristic of large-crab

attack); (iii) sublethal injury; (iv) missing from

tether; or (v) dead without apparent injury. The

response variable was the number of predated and

surviving individuals per replicate, analyzed as bi-

nomial data. Because tethering removes an individ-

ual’s capacity to actively evade predators, absolute

predation proportions from this design should not

be extrapolated as estimates of natural, untethered

predation rates in the eld; instead, tethering was

used to hold baseline vulnerability constant across

treatments and thereby isolate the relative effect of

ALAN, expressed as an odds ratio between ALAN

and control replicates. All analyses were performed

in R 4.3.3 (R Core Team 2024).

Data analysis

A Generalized Linear Mixed Model was tted

to MCH data only using glmmTMB (Brooks et al.

2017). The response was the number of predated

and surviving individuals per replicate (binomi-

al, logit link). Fixed effects were light condition

(ALAN versus natural-darkness control), season

(spring versus summer), and their interaction. Ex-

perimental night was included as a random inter-

cept to account for night-to-night variation. The

need for the interaction term was assessed via like-

lihood-ratio test (LRT) comparing the full model

against an additive model under maximum likeli-

hood. Type-III Wald χ2 tests (car package; Fox and

Weisberg 2019) provided complementary inference.

Table 1. Descriptive summary of tethering experiments at Mar Chiquita (MCH) across two deployment periods and contextual

comparison with the published San Antonio Oeste (SAO) result. Values are mean ± SD across replicates per treatment

per deployment period. MCH data are primary (this study); SAO data are from Nuñez et al. (2021a) and are incorporated

solely in the secondary contextual analysis.

Season Treatment N replicates Total crabs/total predated Mean proportion ± SD

Spring Natural darkness 9 43/27 0.66 ± 0.35

Spring ALAN (≈ 38 lux) 9 45/39 0.87 ± 0.22

Summer Natural darkness 23 113/17 0.15 ± 0.20

Summer ALAN (≈ 38 lux) 23 110/13 0.11 ± 0.18

SAO (published), summer Natural darkness 20 100/22 0.22 ± 0.20

SAO (published), summer ALAN 20 100/69 0.69 ± 0.26

Note: each replicate pools the outcomes of three to ve stakes, each holding one individually tethered juvenile (mean ≈ 4-5 crabs

per replicate); see Tethering experiments.

Marine and Fishery sciences 40 (1): xxx-xxx (2027)

Marginal (R2m) and conditional (R2c) coefcients

were estimated with the performance package (Na-

kagawa and Schielzeth 2013). Residual diagnostics

used DHARMa (Hartig 2018). Pairwise contrasts

(ALAN versus natural darkness within each season)

were computed using emmeans (Lenth et al. 2020)

and are reported as odds ratios (OR) with 95%

condence intervals. Unlike the GLMM, which

appropriately weights each replicate by its own

number of tethered individuals through the bino-

mial structure, the descriptive summary in Table 1

(mean proportion ± SD) is an unweighted average

across replicates, giving each replicate equal weight

regardless of how many individuals it contained.

As a sensitivity check on the choice of random-ef-

fect structure, we additionally tted two block-sen-

sitivity models on the MCH data: one adding block

identity nested within season as a further random in-

tercept alongside night, and one adding it as a xed

effect. Both models supported the same qualitative

conclusions as the primary model (light effect and

light × season interaction both remained signicant,

p < 0.01 in both cases), and the block-random model

was preferred by AIC (ΔAIC = 13.5 relative to the

primary model; likelihood-ratio test for the added

block term: χ21 = 15.46, p < 0.001). We nonetheless

report the primary model (night only) as our main

result because it directly matches the random-effect

structure used in the published SAO reference mod-

el, keeping the three contexts compared under one

consistent specication; the block-sensitivity results

are summarized here to demonstrate that our main

conclusions do not depend on this choice.

To visualise the ALAN-sensitivity gradient

across all three empirical contexts and construct

the forest plot (Figure 3 B), we additionally tted a

GLMM replacing season with a three-level context

factor (MCH summer, MCH spring, SAO summer),

retaining night as a random intercept, and including

the published SAO dataset (Nuñez et al. 2021a)

alongside the primary MCH data. This model was

used solely to generate model-predicted proba-

bilities and odds ratios for all three contexts on

a common parameter scale; all primary inference

is based on the MCH-only model. The published

SAO summer survival OR (ORsurv = 0.12, 95% CI

[0.04-0.35]; Nuñez et al. 2021a) was independently

veried by algebraic inversion (ORpred = 1/0.12 =

8.33, 95% CI [2.86-25.00]).

RESULTS

Descriptive patterns

At MCH, baseline predation (natural-darkness

control) differed markedly between seasons: the

mean predation proportion was 0.66 ± 0.35 (SD)

in spring versus only 0.15 ± 0.20 in summer, re-

ecting the substantial difference in predation pres-

sure between the settlement season and the post-re-

cruitment period (Table 1). Under ALAN, spring

predation rose to 0.87 ± 0.22, whereas summer

predation remained essentially unchanged at 0.11

± 0.18, a divergence that suggests a season-depend-

ent ALAN response already at the descriptive level

(Figure 3 A). At the published SAO site, mean con-

trol predation was 0.22 ± 0.20 and rose to 0.69 ±

0.26 under ALAN (Nuñez et al. 2021a), producing

the highest absolute ALAN effect among the three

contexts examined.

Light × season interaction at MCH

The MCH-only GLMM explained a large pro-

portion of total variation (R2c = 0.884; R2m =

0.591), with the night random intercept absorbing

substantial residual variation (variance = 0.94, SD

= 0.97; ICC ≈ 0.22), conrming that night-to-night

differences warranted their inclusion in the model.

DHARMa residual diagnostics showed no viola-

tions of model assumptions (uniformity p = 0.141;

dispersion p = 0.146; outliers p = 1.000). The LRT

conrmed that the light × season interaction signif-

icantly improved t over the additive model (χ21 =

6.09, p = 0.014, ΔAIC = 4.09). Type-III Wald tests

supported the importance of season (χ21 = 8.99, p

Nuñez et al.: alaN modulates predatioN iN Neohelice graNulata 9

= 0.003) and the interaction (χ21 = 7.42, p = 0.006;

Table 2).

Pairwise contrasts showed that the ALAN effect

was strongly season-dependent (Figure 3 A and B;

Table 2). In spring, ALAN signicantly increased

predation odds relative to natural darkness (OR =

5.77, 95% CI [1.67-19.89], p = 0.006): the odds of

a juvenile being predated were approximately 5.8

times higher under ALAN, with GLMM marginal

probabilities of 0.91 (ALAN) versus 0.64 (control).

In summer, by contrast, ALAN had no detectable

effect on predation probability (OR = 0.75, 95% CI

[0.34-1.65], p = 0.470), with marginal probabilities

of 0.09 (ALAN) and 0.12 (control).

Figure 3. Experimental results for Neohelice granulata at MCH and contextual comparison with SAO. (A) Proportion of juveniles

predated per replicate, by season. Open circles/triangles show individual replicates (circle = Control, triangle = ALAN);

lled symbols show the GLMM-estimated marginal probability ± 95% CI for each treatment, connected by a dashed line.

(B) Forest plot of ALAN effect sizes (odds ratio, log2 scale; dashed line at OR = 1, the null). Points show the estimated

OR and bars the 95% CI. Numeric OR values with 95% CI are given for each context.

MCH (spring)MCH (summer)

ControlALAN ControlALAN

25%

50%

75%

100%

Treatment

Control

ALAN

OR = 8.96

[4.61–17.43]

OR =0.75

[0.34–1.65]

OR = 5.27

[1.60–17.40]

MCH (spring)

This study

MCH (summer)

This study

SAO (summer)

Nuñez et al. (2021)

0.51.0 2.04.0 8.016.0

Odds ratio: ALAN / Control (log 2scale)

Site & season

SAO(summer)

MCH (summer)

MCH (spring)

Proportion of juveniles predated

Marine and Fishery sciences 40 (1): xxx-xxx (2027)

Contextual gradient in ALAN sensitivity

The secondary three-context model conrmed

that the ALAN effect differed signicantly across

contexts (light × context interaction: χ2 = 22.69,

df = 2, p < 0.001; conditional R2 = 0.809, mar-

ginal R2 = 0.654). Effect sizes for each context:

SAO summer OR = 8.96 [4.61-17.43], MCH spring

OR = 5.27 [1.60-17.40], MCH summer OR = 0.75

[0.34-1.65]. These results reveal a clear ordered

gradient in ALAN sensitivity (Figure 3 B; Table

2): SAO summer > MCH spring > MCH summer,

which parallels the seasonal ranking in baseline

(no-ALAN) predation at MCH itself (0.66 in spring

versus 0.15 in summer; Table 1).

DISCUSSION

The seasonal variation in the ALAN effect at

MCH (signicant in spring, entirely absent in sum-

mer) together with the published extreme sensi-

tivity at SAO (Nuñez et al. 2021a), produces an

ordered gradient that is most consistent with the

amplifying hypothesis (HB): the capacity of ALAN

to amplify predation scales with the susceptibility

state of the system (the combined outcome of adult

nutritional condition, juvenile vulnerability and

habitat complexity) rather than with light intensity

alone. Within MCH, where site and protocol are

held constant, this susceptibility state is directly

indexed by seasonal baseline predation itself; the

SAO comparison extends this pattern to an inde-

pendently studied, chronically high-susceptibility

population, though its raw baseline predation value

is not directly comparable across studies (see Cave-

ats and limitations). These results indicate that the

impact of ALAN on predation may be as depend-

ent on the ecological state of the system as on the

intensity of the light itself.

Among the four hypotheses considered, the

observed pattern is not consistent with the com-

pensatory hypothesis (HC), which would predict

the strongest effect at the lowest-susceptibility

context (MCH summer) rather than the weakest,

nor with the additive null (HA), which would pre-

dict a detectable effect even at MCH in summer.

The pattern is also not consistent with a non-lin-

ear threshold model (HD) in its simplest form,

since the ALAN effect continued to increase from

MCH spring to SAO rather than plateauing or

declining at the most susceptible end of the gra-

dient. Taken together, these results are consistent

with studies showing that ALAN affects foraging

most strongly where predators are most motivated

Table 2. Model estimates for ALAN effects on juvenile predation. Primary analysis (MCH-only GLMM): R2m = 0.591, R2c

= 0.884; LRT for light × season: χ21 = 6.09, p = 0.014, ΔAIC = 4.09. DHARMa diagnostics: no violations. Secondary

analysis (three-context GLMM, MCH + SAO): χ22 = 22.69, p < 0.001. OR > 1 indicates ALAN increases predation odds.

Context Model OR 95% CI p/data source

MCH, summer Primary 0.75 [0.34, 1.65] p = 0.470; this study

MCH, spring Primary 5.77 [1.67, 19.89] p = 0.006; this study

MCH, summer Secondary 0.75 [0.34, 1.65] p > 0.05; this study

MCH, spring Secondary 5.27 [1.60, 17.40] p < 0.05; this study

SAO, summer Secondary 8.96 [4.61, 17.43] p < 0.001; Nuñez et al. (2021a)

OR = odds ratio (ALAN/Control); CI = 95% condence interval. The SAO odds ratio used throughout (secondary three-context

model) is independently cross-checked by algebraic inversion of the published SAO survival OR (Nuñez et al. 2021a): ORpred =

1/0.12 = 8.33, 95% CI [2.86-25.00], consistent with the model-based estimate.

Nuñez et al.: alaN modulates predatioN iN Neohelice graNulata 11

(e.g. Becker et al. 2013; Czarnecka et al. 2019)

and that food-web strengthening by ALAN scales

with predator activity levels (Sanders et al. 2018).

It is emphasized, however, that with only three,

non-equivalent contexts, namely two deployment

periods at one site plus a single external, inde-

pendently studied population, our design can rule

out the two hypotheses whose qualitative predic-

tions are clearly violated (HC, HA) but cannot

formally discriminate between the amplifying

(HB) and threshold (HD) hypotheses, nor test

any functional form statistically; doing so would

require a continuous gradient of independently

sampled susceptibility states, which was beyond

the scope of the present study (see Caveats and

limitations below).

Adult nutritional state as a motivational gate-

way for ALAN effects

A plausible explanation for why ALAN ampli-

ed predation at MCH in spring but not in sum-

mer may lie in the difference in adult energetic

state, although we did not measure nutritional

condition directly in the crabs used in this study.

Spring corresponds to the onset of the reproduc-

tive season of N. granulata at MCH (Luppi et

al. 2013) and published hepatosomatic indices

for this population reach their seasonal minimum

in spring (Bas et al. 2014), a nutritional context

that transiently resembles the chronic protein lim-

itation reported for SAO adults year-round (Bas

et al. 2019). If nutritionally stressed adults are

more motivated to exploit any increase in for-

aging opportunity, ALAN could provide exactly

that by extending the period of effective visual

predation on the intertidal surface (see Sztarker

et al. 2020; Nuñez et al. 2021b). This remains a

proposed mechanism consistent with the seasonal

pattern we observed, rather than a directly tested

one; by summer, adults have presumably restored

somatic reserves after reproduction, which could

explain the complete absence of an ALAN effect

during this season.

Juvenile vulnerability and the temporal window

of settlement-season risk

We did not measure megalopa settlement in this

study, so the timing of settlement relative to our

two deployment periods is not resolved by our data.

The juveniles used in our tethering experiments

(12-14 mm CW) were small relative to the adults

(30-45 mm CW) and are likely to be vulnerable to

adult predation: at this size they may still shelter

opportunistically in adult burrow networks rather

than in individual burrows, which would confer

only incomplete refuge from surface-active pred-

ators (Méndez Casariego et al. 2009). Predation

by adults is size-dependent in N. granulata, with

larger juveniles being less vulnerable (Luppi et

al. 2001). At the population level, juveniles re-

cruiting in spring are expected to outgrow the

most vulnerable size classes within a few months,

and by February most individuals present in the

population would have had time to excavate in-

dividual burrows; because summer animals in our

tethering trials were freshly collected rather than

tracked from the spring cohort, this is presented

as a population-level seasonal pattern rather than

a demonstrated trajectory of the same individuals.

The temporal coincidence of maximum juvenile

vulnerability and adult nutritional stress in spring

therefore creates a narrow but critical window

during which MCH’s normally buffering habitat

complexity is partially bypassed.

Habitat complexity as a partial buffer against

ALAN effects

Even within the spring vulnerability window, the

MCH spring ALAN effect (OR = 5.27, from the

secondary three-context model) was lower than

that at SAO (OR = 8.96, same model, with broad-

ly overlapping condence intervals; the published

SAO survival OR is independently cross-checked

in Table 2), which suggests that the dense burrow

network at MCH partially, but not completely,

buffers juveniles even before self-excavation is

Marine and Fishery sciences 40 (1): xxx-xxx (2027)

complete. Habitat complexity appears to act as a

sensitivity oor: the dense burrow network at MCH

reduces the effective foraging range of adult crabs

under ALAN by providing refugia to pre-burrow

juveniles (Méndez Casariego et al. 2009), and the

organic-rich muddy substrate may limit the visual

advantage conferred by illumination compared

with the open sand-cobble ats of SAO. The quan-

titative difference between spring MCH and SAO,

approximately 0.8 units on the log OR scale (a

ratio of ≈ 1.7 between the two odds ratios), could

be attributed to habitat structural complexity as a

third buffering axis, dampening but not eliminat-

ing ALAN-driven predation when the motivational

and vulnerability axes are simultaneously elevated

(Czarnecka et al. 2019).

Demographic consequences and cumulative

population impact

The spring window of ALAN sensitivity at MCH

coincides with the settlement period that contrib-

utes to annual cohort size (Luppi et al. 2013; Bas

et al. 2019). We did not measure population-level

recruitment or perform demographic modelling in

this study, so we cannot quantify the net effect of

ALAN on cohort size. We note, however, a ten-

sion worth agging for future work: Nuñez et al.

(2025) reported that ALAN increased megalopa

settlement at MCH, while the present results show

that a larger proportion of juveniles present during

the spring settlement window are predated under

ALAN. Whether these two effects combine to in-

crease or decrease net recruitment cannot be deter-

mined from either study alone and would require

tracking settlement and post-settlement survival

concurrently.

Caveats and limitations

Several lines of evidence support the robustness

of the spring result despite its narrower replication

history: the light × season interaction remains sig-

nicant whether block is included or excluded from

the model, the effect is consistent in direction and

approximate magnitude across the primary model,

the block-sensitivity models, and the secondary

three-context model, and the resulting OR (5.27-

5.77 depending on model specication) sits within

the ordered gradient observed across all three pred-

atory contexts.

A second limitation concerns predator identity.

At SAO, paired exclusion/inclusion cage experi-

ments identied adult conspecics as the dominant

source of juvenile mortality (Nuñez et al. 2021a);

no equivalent experiment was conducted at MCH

in the present study. Cannibalism by adult conspe-

cics remains the most parsimonious explanation

for the MCH results, given the well-documented

importance of cannibalism in this species (Luppi

et al. 2001), but we cannot exclude a contribution

from other predators (shorebirds or sh, for in-

stance) active during tidal ooding at MCH.

A further limitation is temporal: our comparison

spans a single spring and a single summer at MCH,

rather than multiple years of each season. Inter-an-

nual variation in adult condition, juvenile settle-

ment timing, or predator abundance could all shift

the magnitude of the ALAN effect independently

of season per se, and our design cannot distinguish

a true seasonal effect from a one-off difference be-

tween these particular spring and summer deploy-

ments. Accordingly, throughout this manuscript

‘spring’ and ‘summer’ should be read as labels for

these two specic 2021-2022 deployment periods,

used as an interpretive framework, rather than as

a replicated seasonal factor in the statistical sense.

Finally, and most importantly, with only three

available contexts we can rule out the additive

(HA) and compensatory (HC) hypotheses as sole

explanations, but we cannot formally discrimi-

nate between the amplifying (HB) and non-linear

threshold (HD) hypotheses, nor treat the amplify-

ing hypothesis as statistically conrmed; doing so

would require a larger number of independently

sampled contexts spanning a continuous suscepti-

bility gradient than the three available here, which

was beyond the scope of the present study.

Nuñez et al.: alaN modulates predatioN iN Neohelice graNulata 13

Management implications and concluding re-

marks

The temporal dynamics of ALAN sensitivity doc-

umented here suggest that targeted, rather than uni-

form, lighting restrictions during the juvenile settle-

ment season could be a more efcient conservation

approach than year-round restrictions, though we

did not evaluate the costs or feasibility of any spe-

cic lighting management scheme. Two concrete,

testable interventions follow directly from our

results: (i) shifting ALAN spectral output toward

amber or red, long-wavelength lighting in areas

adjacent to settlement habitat, which is expected to

reduce visual conspicuousness of prey to crustacean

predators relative to cold-white LED (this expecta-

tion should be veried for N. granulata specically

before implementation); and (ii) targeted night-time

lighting curfews strictly conned to the settlement

window once it has been dened locally, rather than

year-round dimming. Neither option has been tested

experimentally here, and both would need site-spe-

cic cost and feasibility assessment. The period of

highest risk at structurally complex intertidal sites

appears to be the juvenile settlement season of N.

granulata, whereas at structurally simpler sites like

SAO, risk may be elevated for longer given the

year-round nutritional stress reported there (Nuñez

et al. 2021a). Season-specic reductions in ALAN

intensity or duration during the settlement window,

where feasible, represent one potential option for

balancing coastal lighting needs against conserva-

tion of intertidal crab populations.

Over two deployment periods at the same site,

the capacity of ALAN to amplify apparent pre-

dation in N. granulata was strongly context-de-

pendent and scaled with the susceptibility state of

the system, a pattern most consistent with, though

not a formal statistical conrmation of, the am-

plifying hypothesis (HB). The ordered gradient in

ALAN sensitivity (SAO summer > MCH spring >

MCH summer) is plausibly explained by the com-

bined severity of adult nutritional stress, juvenile

pre-refuge vulnerability, and habitat openness (fac-

tors we did not measure directly and infer from the

seasonal and site context) rather than by baseline

predation intensity per se. These results indicate

that even high-quality, complex habitats may be

seasonally susceptible to ALAN, that single-season

assessments may underestimate cumulative preda-

tion exposure at the population level (a hypothesis

we could not test directly without demographic

data), and that the conservation value of darkness

is highest, and targeted lighting management most

justied, during the juvenile settlement window.

ACKNOWLEDGEMENTS

We thank N. Chiaradia for aerial photographs

of the experimental setup and logistical support in

the eld, and the staff of IIMyC-UNMdP for infra-

structure and technical assistance. All experimental

procedures complied with Argentine wildlife re-

search regulations (Resolution MAA 264/06 and

subsequent provincial permits).

Ethics and integrity statement

All experimental procedures complied with

Argentine wildlife research regulations (Resolu-

tion MAA 264/06 and subsequent provincial per-

mits) and with institutional guidelines for the use

of animals in research and sampling of intertidal

invertebrates.

Data availability statement

The datasets and code generated during and/

or analyzed during the current study are availa-

ble from the corresponding author on reasonable

request.

Funding declaration

Not applicable.

Marine and Fishery sciences 40 (1): xxx-xxx (2027)

Declaration of competing interests

The authors declare that they have no known

competing nancial interests or personal relation-

ships that could have appeared to inuence the

work reported in this paper.

Declaration of use of Articial Intelligence tools

During the preparation of this manuscript,

the authors used an AI language model (Claude,

Anthropic) solely to improve the English of the

text. The authors reviewed and edited the output

and took full responsibility for the content of the

publication.

Author contributions

Jesús D. Nuñez: methodology design; laboratory

work; eld work; writing original draft; writing-re-

view and editing. Ruben A. Robles: methodology

design; laboratory work; eld work. Macarena

Pérez García: eld work; writing original draft;

writing-review and editing. Camila B. Merodio:

eld work; writing-review and editing. M. Guada-

lupe Vázquez: laboratory work; eld work. Tomás

A. Luppi: methodology design; laboratory work;

eld work; writing-review and editing.

REFERENCES

alberTi J, daleo P, FanJul e, escaPa M, boT-

To F, iribarne o. 2015. Can a single species

challenge paradigms of salt marsh functioning?

Estuaries Coasts. 38: 1178-1188.

aronson rb, heck kl. 1995. Tethering experi-

ments and hypothesis testing in ecology. Mar

Ecol Prog Ser. 121: 307-309.

baker r, WalThaM nJ. 2020. Tethering mobile

aquatic organisms to measure predation: a re-

newed call for caution. J Exp Mar Biol Ecol.

523: 151270.

bas c, lancia JP, luPPi T, Méndez casarieGo

a, kiTTlein M, sPiVak e. 2014. Inuence of

tidal regime, diurnal phase, habitat and season

on feeding of an intertidal crab. Mar Ecol. 35:

319-331.

bas c, nuñez Jd, sPiVak e, luPPi T. 2019. In-

ter-population differences in feeding ecology

under nitrogen-limited diets: eld growth and

the role of cannibalism in limiting/improving

reserve accumulation. Mar Ecol. 40: e12547.

baTTy rs, blaxTer Jhs, richard JM. 1990.

Light intensity and the feeding behaviour of

herring, Clupea harengus. Mar Biol. 107: 383-

388.

becker a, WhiTField ak, coWley Pd, Järne-

Gren J, næsJe TF. 2013. Potential effects of

articial light associated with anthropogenic

infrastructure on the abundance and foraging

behaviour of estuary-associated shes. J Appl

Ecol. 50: 43-50.

bennie J, daVies TW, cruse d, GasTon kJ. 2018.

Articial light at night alters grassland vegeta-

tion species composition and phenology. J Appl

Ecol. 55: 442-450.

bolTon d, Mayer-PinTo M, clark GF, daFForn

ka, brassil Wa, becker a, JohnsTon el.

2017. Coastal urban lighting has ecological

consequences for multiple trophic levels under

the sea. Sci Total Environ. 576: 1-9.

boTTo F, iribarne o. 2000. Contrasting effects of

two burrowing crabs (Chasmagnathus granula-

ta and Uca uruguayensis) on sediment compo-

sition and transport in estuarine environments.

Estuar Coast Shelf Sci. 51: 141-151.

brooks Me, krisTensen k, Van benTheM kJ,

MaGnusson a, berG cW, nielsen a, skauG

hJ, Machler M, bolker bM. 2017. glm-

mTMB balances speed and exibility among

packages for zero-inated generalized linear

mixed modeling. R J. 9: 378-400.

cerri rd. 1983. The effect of light intensity on

predator and prey behaviour in cyprinid sh:

factors that inuence prey risk. Anim Behav.

31: 736-742.

Nuñez et al.: alaN modulates predatioN iN Neohelice graNulata 15

claessen d, de roos aM, Persson l. 2004. Pop-

ulation dynamic theory of size-dependent can-

nibalism. Proc R Soc B. 1537: 333-340.

CzarneCka M, kakareko T, JerMaCz Ł, Paw-

lak r, kobak J. 2019. Combined effects of

nocturnal exposure to articial light and habitat

complexity on sh foraging. Sci Total Environ.

684: 14-22.

daVies TW, bennie J, inGer r, de ibarra nh,

GasTon kJ. 2013. Articial light pollution: are

shifting spectral signatures changing the bal-

ance of species interactions? Glob Chang Biol.

19: 1417-1423.

daVies TW, duFFy JP, bennie J, GasTon kJ. 2014.

The nature, extent, and ecological implications

of marine light pollution. Front Ecol Environ.

12: 347-355.

elGar Ma, cresPi bJ, editors. 1992. Cannibalism:

ecology and evolution among diverse taxa. Ox-

ford: Oxford University Press.

Fox J, WeisberG s. 2019. An R companion to ap-

plied regression. 3rd ed. Thousand Oaks (CA):

Sage.

GarraTT MJ, Jenkins sr, daVies TW. 2019. Map-

ping the consequences of articial light at night

for intertidal ecosystems. Sci Total Environ.

691: 760-768.

GasTon kJ, bennie J, daVies TW, hoPkins J.

2013. The ecological impacts of nighttime light

pollution: a mechanistic appraisal. Biol Rev. 88:

912-927.

GasTon kJ, Visser Me, hölker F. 2015. The bi-

ological impacts of articial light at night: the

research challenge. Philos Trans R Soc B. 370:

20140133.

haMPTon se. 2004. Habitat overlap of enemies:

temporal patterns and the role of spatial com-

plexity. Oecologia. 138: 475-484.

harTiG F. 2018. DHARMa: residual diagnostics

for hierarchical (multi-level/mixed) regression

models. R package version 0.2.0.

hines ah, kraMer Ma, richie kd, sMiTh ld,

aGuilar r, oGburn Mb, Johnson eG, ruiz

GM. 2026. Cannibalism, predation, and refuges

of juvenile blue crabs from a 37-year experi-

ment. PNAS. 123 (12): e2533461123.

hölker F, WolTer c, Perkin ek, Tockner k.

2010. Light pollution as a biodiversity threat.

Trends Ecol Evol. 25: 681-682.

lenTh r, sinGMann h, loVe J, buerkner P,

herVe M. 2020. Emmeans: estimated margin-

al means, aka least-squares means. R package

version 1.5.3.

luPPi Ta, sPiVak ed, anGer k. 2001. Experimen-

tal studies on predation and cannibalism of the

settlers of Chasmagnathus granulata and Cyr-

tograpsus angulatus (Brachyura: Grapsidae). J

Exp Mar Biol Ecol. 265: 29-48.

luPPi Ta, bas cc, Méndez casarieGo Ma, al-

bano M, lancia J, kiTTlein M, rosenThal

a, Farías ne, sPiVak ed, iribarne oo. 2013.

The inuence of habitat, season and tidal re-

gime in the activity of the intertidal crab Ne-

ohelice (=Chasmagnathus) granulata. Helgol

Mar Res. 67: 1-15.

MarTineTTo P, MonTeMayor di, alberTi J, cosTa

csb, iribarne oo. 2016. Crab bioturbation

and herbivory may account for variability in

carbon sequestration and stocks in south west

Atlantic salt marshes. Front Mar Sci. 3: 12.

McMahon o, sMyTh T, daVies TW. 2022. Broad

spectrum articial light at night increases the con-

spicuousness of camouaged prey. J Appl Ecol.

59: 1324-1333. DOI: https://doi.org/10.1111/13

65-2664.14146

Méndez casarieGo a, alberTi J, luPPi Ta, iri-

barne oo. 2009. Stage-dependent interactions

between intertidal crabs: from facilitation to

predation. J Mar Biol Assoc UK. 89: 781-788.

Méndez casarieGo a, alberTi J, luPPi Ta,

daleo P, iribarne oo. 2011. Habitat shifts

and spatial distribution of the intertidal crab

Neohelice (Chasmagnathus) granulata Dana. J

Sea Res. 66: 87-94.

Moksnes Po. 2004. Self-regulating mechanisms

in cannibalistic populations of juvenile shore

crabs Carcinus maenas. Ecology. 85: 1343-

1354.

Marine and Fishery sciences 40 (1): xxx-xxx (2027)

Moyse e, FirTh lb, sMyTh TJ, Tidau s, daVies

TW. 2023. Articial light at night alters preda-

tion on colour-polymorphic camouaged prey.

Basic Appl Ecol. 73: 88-93.

nakaGaWa s, schielzeTh h. 2013. A general and

simple method for obtaining R2 from general-

ized linear mixed-effects models. Methods Ecol

Evol. 4: 133-142.

nuñez Jd, bas c, Pérez García M, ocaMPo eh,

ribeiro Pd, luPPi Ta. 2021a. Articial light at

night may increase the predation pressure in a

salt marsh keystone species. Mar Environ Res.

167: 105285.

nuñez Jd, Pérez García M, Merodio cb, luPPi

Ta. 2025. The effect of articial light at night

on settlement patterns of a salt marshes crab.

BMC Environ Sci. 2: 23.

nuñez Jd, sbraGaGlia V, sPiVak ed, chiaradia

nM, luPPi Ta. 2021b. The magnitude of be-

havioural responses to articial light at night

depends on the ecological context in a coastal

marine ecosystem engineer. Mar Environ Res.

165: 105238.

Perkin ek, hölker F, richardson Js, sadler JP,

WolTer c, Tockner k. 2011. The inuence of

articial light on stream and riparian ecosys-

tems: questions, challenges, and perspectives.

Ecosphere. 2: art122.

Polis Ga. 1981. The evolution and dynamics of

intraspecic predation. Annu Rev Ecol Syst. 12:

225-251.

r core TeaM. 2024. R: a language and envi-

ronment for statistical computing. Vienna: R

Foundation for Statistical Computing. https://

www.R-project.org/.

sanders d, kehoe r, cruse d, Van Veen FFJ,

GasTon kJ. 2018. Low levels of articial light

at night strengthen top-down control in insect

food web. Curr Biol. 28: 2474-2478.e3.

sPiVak e, anGer k, luPPi T, bas c, isMael d.

1994. Distribution and habitat preferences of

two grapsid crab species in Mar Chiquita La-

goon (Province of Buenos Aires, Argentina).

Helgol Meeresunters. 48: 59-78.

szTarker J, berón de asTrada M, ToMsic d.

2020. The visual system of Neohelice: from gen-

eral structure to cytoarchitecture. In: rodríGuez

eM, luPPi Ta, editors. Neohelice granulata, a

model species for studies on crustaceans. Vol-

ume II: Anatomy and physiology. Cambridge

Scholars Publishing, Newcastle upon Tyne. p.

29-50.

underWood cn, daVies TW, queirós aM. 2017.

Articial light at night alters trophic interac-

tions of intertidal invertebrates. J Anim Ecol.

86: 781-789.

zaPaTa MJ, sulliVan sMP, Gray sM. 2019. Ar-

ticial lighting at night in estuaries-implica-

tions from individuals to ecosystems. Estuaries

Coasts. 42: 309-330.