Received: 23 April 2022 | Revised: 16 May 2022 | Accepted: 17 May 2022

DOI: 10.1002/ece3.8990

RESEARCH ARTICLE

Patterns and drivers of Nothobranchius killifish diversity in

lowland Tanzania

Martin Reichard1,2,3 | Michal Janáč1,3 | Radim Blažek1,3 | Jakub Žák1,4 |

Okinyi David Alila5,6 | Matej Polačik1

1
Institute of Vertebrate Biology, Academy
of Sciences of the Czech Republic, Brno, Abstract
Czech Republic
Temporary pools are seasonal wetland habitats with specifically adapted biota, in-
2
Department of Ecology and Vertebrate
Zoology, University of Łódź, Łódź, Poland
cluding annual Nothobranchius killifishes that survive habitat desiccation as diapaus-
3
Department of Botany and Zoology, ing eggs encased in dry sediment. To understand the patterns in the structure of
Faculty of Science, Masaryk University, Nothobranchius assemblages and their potential in wetland conservation, we com-
Brno, Czech Republic
4 pared biodiversity components (alpha, beta, and gamma) between regions and esti-
Department of Zoology, Faculty of
Science, Charles University, Prague, Czech mated the role and sources of nestedness and turnover on their diversity. We sampled
Republic
5
Nothobranchius assemblages from 127 pools across seven local regions in lowland
Institute of Ecology and Evolution,
University of Bern, Bern, Switzerland Eastern Tanzania over 2 years, using dip net and seine nets. We estimated species
6
Department of Fish Ecology and composition and richness for each pool, and beta and gamma diversity for each re-
Evolution Centre of Ecology, Evolution
gion. We decomposed beta diversity into nestedness and turnover components. We
and Biogeochemistry, EAWAG Swiss
Federal Institute of Aquatic Science and tested nestedness in three main regions (Ruvu, Rufiji, and Mbezi) using the number of
Technology, Kastanienbaum, Switzerland
decreasing fills metric and compared the roles of pool area, isolation, and altitude on
Correspondence nestedness. A total of 15 species formed assemblages containing 1–­6 species. Most
Martin Reichard, Institute of Vertebrate
Nothobranchius species were endemic to one or two adjacent regions. Regional diver-
Biology, Academy of Sciences of the
Czech Republic, Brno, Czech Republic. sity was highest in the Ruvu, Rufiji, and Mbezi regions. Nestedness was significant in
Email: [email protected]
Ruvu and Rufiji, with shared core (N. melanospilus, N. eggersi, and N. janpapi) and com-
Funding information mon (N. ocellatus and N. annectens) species, and distinctive rare species. Nestedness
Grantová Agentura České Republiky,
apparently resulted from selective colonization rather than selective extinction, and
Grant/Award Number: 19-­01781S
local species richness was negatively associated with altitude. The Nothobranchius as-
semblages in the Mbezi region were not nested, and had many endemic species and
the highest beta diversity driven by species turnover. Overall, we found unexpected
local variation in the sources of beta diversity (nestedness and turnover) within the
study area. The Mbezi region contained the highest diversity and many endemic spe-
cies, apparently due to repeated colonizations of the region rather than local diversifi-
cation. We suggest that annual killifish can serve as a flagship taxon for small wetland
conservation.

KEYWORDS
Africa, Cyprinodontiformes, dispersal, ephemeral habitats, habitat protection

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium,
provided the original work is properly cited.
© 2022 The Authors. Ecology and Evolution published by John Wiley & Sons Ltd.

Ecology and Evolution. 2022;12:e8990.  www.ecolevol.org | 1 of 12

https://doi.org/10.1002/ece3.8990
2 of 12 | REICHARD et al.

TA X O N O M Y C L A S S I F I C AT I O N
Biogeography; Community ecology; Conservation ecology

1 | I NTRO D U C TI O N The conservation status of all 94 currently recognized

Nothobranchius species has been recently evaluated (Nagy &
Freshwater wetlands represent the interface between aquatic and Watters, 2021). Nothobranchius fishes are well characterized taxo-
terrestrial habitats, uniquely combining species of both ecosystems. nomically and contain widespread species and species with locally
Their extent and broader ecological significance vary seasonally, endemic distribution, ranging from species of least concern to crit-
with annual fluctuation of environmental conditions associated with ically endangered (Nagy & Watters, 2021). Several Nothobranchius
relatively humid and drier periods. Consequently, only a few spe- fishes often coexist, forming local assemblages with up to six species
cies are primarily associated with wetlands and adapted to their dy- in a single temporary pool. Tanzania is the distribution center of the
namic environment (Dalu & Wasserman, 2022; van der Valk, 2012). genus, with 47 of the 94 currently known Nothobranchius species in
Wetlands are also a source of immediate and long-­term benefits to Africa, of which many have restricted ranges. The highest species
human societies, from fishing and livestock support to irrigation and richness is found in coastal Tanzania (van der Merwe et al., 2021).
recession agriculture, although many have been seriously degraded Temporary pools in this region are considerably affected by habitat
(Dalu et al., 2017). degradation and loss to human activities, such as urbanization and
Temporary pools are a common feature of freshwater wetlands agriculture expansion (Nagy & Watters, 2021).
(Williams, 2006). They are seasonal habitats and aquatic species may To evaluate their possible role as a flagship taxon in temporary
either recolonize each pool during the wet season, or be adapted to wetland conservation and understand the general patterns in the
persist throughout the wet and dry phases of the annual cycle (Dalu & structure of Nothobranchius assemblages and their biodiversity, we
Wasserman, 2022). Annual recolonization of temporary pools is typ- investigated alpha (species richness in a local pool), beta (composi-
ical for amphibians, migratory fishes, and flying insects, while local tional differences between local pools within a region), and gamma
persistence is characteristic of specialized invertebrate taxa, such (regional species richness) diversity and their drivers in 127 tem-
as large branchiopods, with local egg banks surviving unfavorable porary pools from seven regions in lowland Tanzania. Specifically,
conditions as a resting stage (Brock et al., 2003; Vanschoenwinkel we (1) compared local (alpha) and regional (gamma) diversity of
et al., 2011). Temporary pools, and especially their resident biota, Nothobranchius assemblages in particular regions (defined by river
are ideal candidates for environmental assessment of wetland con- catchments) and identified the locally most common species, (2)
ditions (Freiry et al., 2020). estimated the level of nestedness in three major regions, (3) tested
Nothobranchius is a genus of small freshwater annual killifish of the whether the main source of nestedness was selective extinction or
family Nothobranchiidae (order Cyprinodontiformes; Wildekamp, selective colonization, and (4) compared the roles of nestedness and
2004). The genus is composed of short-­living (3–­12 months), sexually species turnover on beta diversity in the three main regions.
dimorphic species (Wildekamp, 2004). Most species reach only 30–­
70 mm in standard length, with one lineage (containing two species)
reaching 10 cm or more (Lambert et al., 2019). Nothobranchius fishes 2 | M E TH O D S
are obligatory wetland species, associated with temporary pools
(Wildekamp, 2004). Unusually for fish, their life history combines 2.1 | Fish sampling
enduring resting stages that survive desiccation, with periods of
post-­hatching aquatic life (Reichard & Polačik, 2019). Nothobranchius All data were collected during two field expeditions to lowland
embryos survive the dry part of the annual cycle in a diapausing Eastern Tanzania. The study area (Figure 1) encompassed the major
stage, with minimal metabolism and high capacity to withstand en- part of the Coastal ichthyological province (sensu Lévêque, 1997),
vironmental challenges (Polačik et al., 2021). Shortly after habitat which is formed by the basins of the East African rivers flowing into
inundation, the eggs hatch and Nothobranchius fish grow to sexual the Indian Ocean between the Nilo-­Soudanian region in the north
maturity within a few weeks (Vrtílek et al., 2018). Nothobranchius and the Zambezi system in the south (Roberts, 1975). The sampling
fishes are weak competitors (Reichard, 2015). They are not capable was conducted from May 25 to June 7, 2017, and May 25 to June
of enduring prolonged competition with fluvial fishes. Their popula- 11, 2019, at the peak abundance of Nothobranchius fishes in the final
tions are often constrained to isolated (typically rainwater-­fed) pools stage of the main rainy season. The sampling region has two rainy
(Reichard, 2015). Annual killifish dispersal is limited and appears to seasons, the main rainy season with more intensive precipitation
be a combination of occasional large-­scale flooding and geomorpho- lasts from March to May and a shorter one from November to mid-­
logical changes to the river basins (Bartáková et al., 2015, 2020; van December (Watters et al., 2020). During sampling, we targeted sites
der Merwe et al., 2021). with the supposed presence of Nothobranchius fishes. In particular,
REICHARD et al. | 3 of 12

F I G U R E 1 Map of the study area,

with sampled pools indicated by dots
and grouped into individual regions.
Many dots were jittered to enable their
visualization, and their GPS coordinates
are provided in associated Dryad file
(10.5061/dryad.m63xsj44f). Colors
denote species richness (black: 1, dark
blue: 2, light blue: 3, orange: 4, red: 5 or 6
species)

sites were selected on the basis of at least assumed temporal isola- measured the shortest distance to the main regional river. Fish were
tion from the main flowing waters. Many of those pools were rela- collected using a dip net with a triangular metal frame (45 × 45 cm)
tively distant from a stream and were apparently formed by locally and 5 mm mesh size, with a 1.5 m wooden pole. Typically, 15 to 80
drained rainwater rather than inundation from a stream, including dip-­net hauls were swept at each site but fewer hauls (at least five)
pools located in roadside ditches and culverts (which typically rep- were performed in the smallest sites where five hauls already col-
resented former natural depressions). Some pools were likely filled lected most fish individuals from the pool. In larger pools, we used a
with river water during flooding and left isolated after water reces- seine net (length 2.7 m, depth 0.7 m, mesh size 4 mm) in addition to
sion and a few pools still had connections to minor streams. Some dip-­net sampling. The mesh size used retained adult Nothobranchius
sites were active rice paddies. unselectively and there was no apparent species-­specific bias in the
We collected 185 samples (83 sites in 2017 and 102 sites in 2019). probability of capture (Reichard et al., 2014). Fish were typically un-
A total of 57 samples (30% of sites) contained no Nothobranchius and ambiguously identified to species in the field and most individuals
were not considered in the analyses. One site was found to be lo- were returned to the pools. Voucher individuals, or individuals that
cated outside the defined study area and was deleted from the final needed genetic confirmation of species identification (e.g., isolated
dataset. The final dataset (127 assemblages) contained seven sites females which are more difficult to assign to a species), were taken
with identical geographic position sampled in both years. We consid- to the laboratory.
ered those samples as independent sampling units, as wetlands form Sampling was carried out in accordance with relevant local
a large inundated area and isolated pools are formed during water and international guidelines and regulations. Sample collection
recession. Indeed, only in one of these sites were the Nothobranchius complied with legal regulations of Tanzania (research permit:
assemblages identical between the two sampling years. At each RPGS/R/AS/11/2017) and export of voucher individuals was ap-
site, we recorded GPS location and altitude (Garmin nüvi 550). We proved through research associateships with Sokoine University of
estimated pool size and measured its maximum depth to the near- Agriculture in Morogoro, Tanzania, and associated permit AS/A/1.
est 5 cm. We used Google Earth software to assign each site to a All data were lawfully acquired in accordance with The Nagoya
particular region based on local topography upon reference to a Protocol on Access to Genetic Resources and the Fair and Equitable
high-­resolution map (Garmin and Google Earth satellite images) and Sharing of Benefits Arising from Their Utilization to the Convention
4 of 12 | REICHARD et al.

on Biological Diversity, with the granted exemption in our case, as nestedness (Ulrich, 2009). We considered alpha level of 0.05 for sta-
the legal procedures to obtain appropriate permissions were not fi- tistical significance.
nalized and implemented in the country of data collection (Tanzania) Sorting the matrices along the gradient of altitude was further
in the years 2017 and 2019, and our research used in situ species added to our analyses because our experience with the study taxon
determination and release of live individuals. (Reichard et al., 2017) suggests that species-­specific responses to
this parameter can drive nested patterns more than pool area and
isolation. As the matrices were ordered along the vertical axis, we
2.2 | Data analysis searched only for the nestedness of sites, using Nrows value of NODF
(Almeida-­Neto et al., 2008). Each of the ordered matrices within
For each site, we tabulated the presence/absence of each each system was compared to 999 matrices originating by randomly
Nothobranchius species and associated geographic and environmen- “shuffling” rows of the original matrix. The nestedness analysis was
tal variables. For each region, we calculated the mean alpha diver- conducted using R 3.5.2 (R Core Team, 2018), using vegan library
sity (mean number of species in a pool, across all sites with at least (Oksanen et al., 2013).
one Nothobranchius species within a particular region), beta diver-
sity (change in assemblage composition across sites, calculated ac-
cording to Baselga, 2010), and gamma diversity (sum of all species 3 | R E S U LT S
recorded within the region). We then further tested how the assem-
blages were structured. 3.1 | Frequency of species occurrence, shared
To determine nestedness in each of the three main regions (Ruvu, species, and endemism
Rufiji, and Mbezi; selected for their sufficient number of sampled as-
semblages), we used the number of decreasing fills (NODF, Almeida-­ Overall, 127 of the sampled pools contained at least one
Neto et al., 2008) as a nestedness metric. This index can reach values Nothobranchius species (Figure 1). The Wami region (12 pools
from 0 (completely random) to 100 (perfectly nested). It was sug- with Nothobranchius present) contained only a single species (the
gested as an alternative to metrics such as matrix temperature (Atmar widespread N. melanospilus). Nothobranchius pools in the regions
& Patterson, 1993) or discrepancy measure (Brualdi & Sanderson, Kilombero (seven pools, one to three species), Northern coastal
1999), with NODF corresponding more properly to a definition of nest- floodplain (four pools, one to two species), and Pangani (two pools,
edness as a system organization, in which less rich sites/species are one species in each) were too scarcely represented in our sampling
subsets of richer sites/species (Almeida-­Neto et al., 2008). Statistical to include in further analyses. Nothobranchius assemblages in the
significance was tested by comparing the NODF values with 999 ran- other three regions (Ruvu: 47 pools, Rufiji: 36 pools, and Mbezi: 16
dom assemblages created according to a null model. We chose the r1 pools) were further investigated for their structure, nestedness,
null model (corresponding to RANDOM1 constraints of Patterson & beta diversity, and their sources. We detected one to six species co-
Atmar, 1986), in which row totals (number of species at a site) remain existing in a pool, with each of the three main regions harboring at
fixed, with species being “drawn” with probabilities weighted by their least two sites with four Nothobranchius species. Local species rich-
observed incidence values (Wright et al., 1997). This model was cho- ness (alpha diversity) was comparable across the three regions and in
sen as a compromise between the equiprobable–­equiprobable model, the Kilombero (Table 1). Higher species richness was found in pools
which is truly random, but too prone to type I errors (Ulrich, 2009; located lower than 30 m above sea level (masl; Figure 2).
Wright et al., 1997) and fixed-­fixed models that preserve both margin The Ruvu and Rufiji shared the most species, with N. melanospi-
totals, which risks being too conservative (Almeida-­Neto et al., 2008). lus present in almost all pools containing Nothobranchius (92.2% and
As NODF, similarly to other nestedness metrics, changes with matrix 91.6% in the Ruvu and Rufiji, respectively). Nothobranchius eggersi
fill (i.e., proportion of 1s in the matrix; Almeida-­Neto et al., 2008) and (54.9% and 27.8%) and N. janpapi (47.1% and 47.2%) were also com-
standardized effect sizes were used to compare nestedness between mon species, followed by predatory N. ocellatus (17.6% and 16.7%)
the three systems studied (Ulrich, 2009). and N. annectens (11.8% and 16.7%). The rarest species (found only
In the significantly nested systems, we compared the nestedness in two pools in the regions) differed between the Ruvu (N. foerschi)
of sites (i.e. rows), between matrices sorted along gradients of pool and Rufiji (N. lourensi). The Mbezi region harbored a relatively differ-
area, isolation (measured as a distance from the main river stem), ent and more diverse assemblage of Nothobranchius species, with N.
and altitude. Ordering site-­species matrices of the nested systems ruudwildekampi (43.8%), N. melanospilus (37.5%), and N. luekei (37.5%)
along gradients of patch area and degree of isolation can help deter- the most frequent species. In the Kilombero region, N. lourensi oc-
mine whether the system is colonization or extinction driven (Ulrich, curred in most (85.7%) Nothobranchius pools. The Mbezi region had
2009). According to Ulrich (2009), extinction-­driven system matri- five endemic species. Kilombero had one endemic species and both
ces would be nested along the gradient of area but not isolation, as Pangani species were endemic to that region (Table 1). The num-
the colonization in these systems would not be strong enough to ber of endangered species was highest in the Mbezi region (three of
generate nestedness. Lack of nestedness in the area-­sorted matrix seven species) and all Nothobranchius species in the Kilombero and
then would result in rejecting the hypothesis of mass-­effect-­driven Pangani regions were classified as vulnerable (Table 1).
REICHARD et al. | 5 of 12

TA B L E 1 List of collected species, with their IUCN conservations status (Nagy & Watters, 2021) and the number of populations detected
in particular regions (with the number of Nothobranchius assemblages investigated in a region in parentheses)

Ruvu Rufiji Mbezi Kilombero Wami North Pangani

Species IUCN (n = 51) (n = 36) (n = 16) (n = 7) (n = 12) (n = 4) (n = 2)

N. albimarginatus En –­ –­ 4 –­ –­ –­ –­
N. annectens NT 6 6 –­ –­ –­ –­ –­
N. eggersi LC 28 10 –­ –­ –­ 3 –­
N. foerschi Vu 2 –­ –­ –­ –­ –­ –­
N. geminus Vu –­ –­ –­ 3 –­ –­ –­
N. janpapi LC 24 17 –­ –­ –­ –­ –­
N. lourensi NT –­ 2 2 6 –­ –­ –­
N. lucius NT –­ –­ 1 5 –­ –­ –­
N. luekei En –­ –­ 6 –­ –­ –­ –­
N. melanospilus LC 47 33 6 –­ 12 4 –­
N. ocellatus NT 9 6 –­ –­ –­ –­ –­
N. palmqvisti Vu –­ –­ –­ –­ –­ –­ 1
N. rubripinnis EN –­ –­ 5 –­ –­ –­ –­
N. ruudwildekampi Vu –­ –­ 7 –­ –­ –­ –­
N. vosseleri Vu –­ –­ –­ –­ –­ –­ 1
Alpha diversity –­ 2.28 2.00 2.00 2.00 1.00 1.75 1.00
Beta diversity –­ 0.77 0.79 0.87 –­ –­ –­ –­
Gamma diversity –­ 6 6 7 3 2 2 2
Vu + En 2 of 6 0 of 6 6 of 7 3 of 3 0 of 2 0 of 2 2 of 2

Note: For reach region, measures of alpha, beta, and gamma diversity are provided, as well as the sum of species in “Vulnerable” and “Endangered”
IUCN categories combined (En: endangered, NT: near threatened, LC: least concern, Vu: vulnerable). Beta diversity was only calculated in three
regions with sufficient number of investigated assemblages. Note that we failed to collect two critically endangered species and two vulnerable
species reported from the study area.

3.2 | Nestedness assemblage nestedness. Its effect was statistically significant in the
Ruvu region (Nrows = 43.90, p = .018), but not in the Rufiji region
The Ruvu and Rufiji regions contained significantly nested (Nrows = 33.65, p = .104).
Nothobranchius assemblages (Ruvu: NODF = 69.68; and Rufiji:
NODF = 55.41, Figure 3) –­ their observed NODF differed from
randomized matrices (both p = .001; Table 2). In contrast, the 3.4 | The role of nestedness and turnover on
Nothobranchius assemblages in the Mbezi region were not nested beta diversity
(NODF = 33.89, which was not significantly different from rand-
omized matrices; p = .766). The standardized effect sizes of NODF The highest beta diversity was found in the Mbezi region, strongly
were 4.82, 2.97, and −0.75 for the Ruvu, Rufiji, and Mbezi regions, driven by species turnover (Figure 5). The Rufiji and Ruvu regions
respectively. had similar beta diversity (Figure 5a), but it was primarily driven by
nestedness in the Ruvu region and by species turnover in the Rufiji
(Figure 5b,c).
3.3 | The sources of nestedness

Nestedness was not driven by selective extinction because area-­ 4 | DISCUSSION

sorted matrices were not associated with nestedness in any re-
gion (Ruvu: Nrows = 36.43, p = .364; Rufiji: Nrows = 28.68, p = .422; Nothobranchius are a specialized fish lineage inhabiting temporary
Figure 4). Selective colonization appeared to have at least a mar- pools in African wetlands. We described the patterns of local di-
ginal effect on nestedness because the isolation-­sorted matrix had versity and assemblage structure in the area where their diversity
a minor effect on nestedness in the Ruvu region (Nrows = 41.75, is globally the highest –­ lowland Tanzania. We identified region
p = .054), although not in the Rufiji region (Nrows = 28.52, p = .431). (Mbezi Triangle) where local endemism was the highest and beta di-
Altitude was found to be the most important factor associated with versity was driven by species turnover. In the two spatially largest
6 of 12 | REICHARD et al.

F I G U R E 2 Relationship between
Nothobranchius species richness and
altitude in (a) the Ruvu and (b) Rufiji
regions

regions (the basins of the lower Rufiji and lower Ruvu rivers), we Luhule, and separated from the Ruvu and Rufiji basins by the Pugu
found that Nothobranchius assemblages were nested, apparently and Mtoti Hills) harbored five endemic species (71% of local spe-
due to selective colonization rather than selective extinction and cies). The Kilombero and Pangani regions (with two of three, and
local species richness was negatively associated with altitude. Most two of two Nothobranchius species being endemic to the region, re-
Nothobranchius species were endemic to one or two adjacent re- spectively) were also composed of locally isolated species, although
gions (Table 1), with a single species, N. melanospilus, occurring with lower species richness. The Ruvu and Rufiji regions shared the
across most of the study area. major part of their Nothobranchius species. These two regions are
connected via a swampy area in their middle and upper reaches and
are known to share mitochondrial lineages of the widespread N. mel-
4.1 | Regional endemism and species distributions anospilus (Bartáková et al., 2020).
N. melanospilus was the most common species in the study area,
Our study area formed a major part of the Coastal ichthyological pertaining to all but two regions. This species is locally common
province (sensu Lévêque, 1997) in East Africa, which is composed and often represents the only Nothobranchius species in temporary
of river basins flowing into the Indian Ocean north of the Zambezi pools, especially in the Ruvu, Rufiji, and Wami regions. Its exclusive
(Roberts, 1975). We identified relatively strong Nothobranchius en- presence was more common in, although not restricted to, higher
demism in several regions of this province. In particular, the Mbezi altitude sites. It is a generalist species (Wildekamp, 2004) with per-
Triangle (composed of minor coastal basins of the Karole, Mbezi, and haps more effective dispersal ability. Nothobranchius melanospilus
REICHARD et al. | 7 of 12

F I G U R E 3 Occurrence of Nothobranchius species in three Tanzanian coastal regions visualized as incidence in a packed matrix (i.e.,
species–­sites matrix sorted by species richness and frequency, representing species-­by-­site matrix with minimal nestedness temperature).
The NODF values are presented in the bottom right corner of each panel

has been recently divided into three cryptic species (Costa, 2019). contained other Nothobranchius species. The other two local spe-
While populations assigned to N. kwalensis occur outside our study cies not recorded in our samples were N. flammicomantis (endemic
area, N. prognathus was described from the Wami basin. We treat to the Ruvu) and N. kilomberoensis (endemic to the Kilombero) –­ both
N. melanospilus as a single species because the validity of the cryp- species are not widespread and are classified as vulnerable (Nagy &
tic species has been questioned (Bartáková et al., 2020; Wildekamp, Watters, 2021), and we may have missed them in our sampling due
2019) and is not widely recognized (van der Merwe et al., 2021). If to interannual variation in their population densities (Nagy & Kiss,
we were to accept the split of N. melanospilus, the only change in 2010; Shidlovsky, 2010).
our conclusion would be that N. prognathus (rather than N. melano-
spilus) occurs in the Wami region, and parts of the upper section of
the Ruvu region. All other regions apparently contain N. melanospilus 4.2 | Local assemblages and sources of their
sensu stricto (Bartáková et al., 2020) and all our main conclusions on structure
assemblage structure would remain unchanged. We acknowledge
that a thorough evaluation of Nothobranchius diversity on genome-­ Patterns of beta diversity, and its partitioning into two main
scale data is needed. components –­ species turnover and nestedness –­, provide an
The concept of biological communities posits that some species important insight into the structuring of biological communities
are naturally very rare (Magurran, 2003), either through their special (Baselga, 2010; Socolar et al., 2016). Local assemblages are shaped
ecological requirements or particular evolutionary history. We failed by interactions between local and regional processes, which can
to collect four species known from our study area. Two species, N. be neutral (random subsets of regional metacommunity) or de-
fuscotaeniatus and N. steinforti, are classified as critically endangered, terministic (species sorting according to local conditions; García-­
known only from their type localities, and collected only in the years Navas et al., 2022). Hence, species turnover and nestedness
of their discovery (1997 and 1976, respectively; Nagy & Watters, reflect different mechanisms underlying the change in assemblage
2021). Our sampling scheme included their type localities and sur- structure across the region –­ species replacement or their selec-
rounding wetlands, which appear natural and uncompromised, and tive loss (Bevilacqua & Terlizzi, 2020).
8 of 12 | REICHARD et al.

TA B L E 2 Nestedness parameters of Nothobranchius assemblages of the Zhoushan archipelago was related to selective extinctions in
in (a) Ruvu, (b) Rufiji, and (c) Mbezi regions (matrix fill 0.372, 0.362, smaller islands (Xu et al., 2017). Our analysis did not support the
and 0.187, respectively)
hypothesis that Nothobranchius assemblage nestedness was driven
NM NM by selective extinctions. Instead, our data are consistent with the
Metric Value 50% 95% p SES scenario of selective colonization leading to assemblage nestedness.
(a) Ruvu Ephemeral wetland pools are formed within existing floodplain and
Ncolumns 64.93 56.44 64.77 .045 1.58 frequently colonized from adjacent habitat patches at relatively

Nrows 69.74 55.94 60.81 .001 4.84 short time scales. In contrast, Zhoushan is a land bridge archipelago,
separated from the mainland only 7–­9 kya, and local species richness
NODF 69.68 55.92 60.73 .001 4.82
is likely driven by selective extinctions in originally species-­rich as-
(b) Rufiji
semblages (Xu et al., 2017) over considerably longer time scales than
Ncolumns 54.71 44.27 52.98 .021 1.98
in the wetland ecosystem.
Nrows 55.42 44.74 50.98 .005 2.95
Strong assemblage nestedness reinforces the division into few
NODF 55.41 44.72 50.91 .005 2.97
common and many rare species. Rare species are intrinsically vul-
(c) Mbezi nerable and often become endangered (Loiseau et al., 2020). Species
Ncolumns 39.44 37.14 51.19 .389 0.24 rareness may be a function of their evolutionary history (Pepke
Nrows 32.92 36.81 44.60 .820 −0.91 et al., 2019) or occupation of a specialized niche (Dorado et al., 2011;
NODF 33.89 36.68 44.67 .766 −0.75 Markham, 2014). Evolutionary causes of rareness in Nothobranchius
Note: Metric values calculated for nestedness among species (Ncolumns), have been hypothesized in the context of range expansion and con-
sites (Nrows), and for the whole system are presented, along with their traction effects as a function of evolutionary age (van der Merwe
standardized effect sizes (SES), medians, and 95% quantiles of the et al., 2021). An example of an ecological source of rareness is the
values calculated for 999 simulated randomized matrices (NM 50%, NM
case of a highly derived trophic niche in N. ocellatus. This species is a
95%) and the probability (P) that the calculated value exceeds the range
of simulated matrices. unique predator of other Nothobranchius species in the community
(Lambert et al., 2019; Watters et al., 2020) and reaches up to 12 cm,
two-­ to three-­fold larger than coexisting Nothobranchius species. It is
The highest beta diversity was found in the Mbezi region and typically recorded in assemblages with four Nothobranchius species
arose almost exclusively through species turnover. In contrast, (range 3–­5) and apparently requires the presence of congeners that
beta diversity in the Ruvu Nothobranchius assemblages was driven serve as its prey (Wildekamp, 2004). Similar evolution of giant pred-
by nestedness. In the Rufiji region, turnover and nestedness were atory species in annual fish assemblages has been reported from the
of intermediate importance. This unexpected local variation in the Neotropics (Helmstetter et al., 2020) and demonstrates that wetland
sources of beta diversity demonstrates that even within the same killifish assemblages from various regions converge to a common
taxon and spatially constrained study area, the sources of beta diver- pattern (García et al., 2009, 2019; Loureiro et al., 2015) through ad-
sity can vary substantially. The Mbezi region is most diverse topo- aptations to specialized ecological niches. While the body morphol-
graphically and is composed of small coastal basins interspersed by ogy of Nothobranchius species is relatively conserved (Wildekamp,
hills and ridges. While this could make local isolation and evolution 2004) and species evolve toward three major body size optima
of locally endemic species more likely than in the larger floodplains (Lambert et al., 2019), there is preliminary evidence of trophic niche
of the Ruvu and Rufiji basins, the phylogenetic hypotheses of the (Polačik et al., 2014; Polačik & Reichard, 2010) and trophic-­related
genus (van der Merwe et al., 2021) reject the existence of a locally morphology (Costa, 2018; Reichard, 2015) differentiation within
endemic species flock in the Mbezi region. It appears that repeated local Nothobranchius assemblages. A strong phylogenetic hypothesis
colonizations of this region, and perhaps independent speciation and comprehensive knowledge of the ecological niche of particular
events, resulted in the highest local species richness and diversity species are needed to understand the role of evolutionary processes
within the entire Nothobranchius range (Lambert et al., 2019; van der in structuring Nothobranchius assemblages.
Merwe et al., 2021). This can facilitate rapid species turnover at a
small geographic scale.
Nestedness describes positive species co-­occurrence result- 4.3 | Conservation
ing in species-­poorer local assemblages being predictable subsets
of species-­richer local assemblages (Almeida-­Neto et al., 2008). The understanding of how nestedness and turnover are linked to
Nothobranchius assemblages were clearly nested in the Ruvu and beta diversity is important for defining optimal conservation strate-
Rufiji regions. Nested patterns of local assemblages are primarily gies (Socolar et al., 2016). The positive correlation of species turn-
driven by extinction–­colonization dynamics (rather than dispersal over with beta diversity is common in tropical assemblages and
and local environmental conditions), and are typically associated associated with habitat heterogeneity (Soininen et al., 2018). This
with larger spatial scales and higher latitudes (Soininen et al., 2018). pattern is congruent across coexisting functional and taxonomic
For example, significant nestedness in local butterfly assemblages groups (Gibson et al., 2017), making Nothobranchius assemblages
REICHARD et al. | 9 of 12

F I G U R E 4 Nestedness of sites (Nrows)

for 999 randomly ordered matrices,
their median (large red point), and 95%
quantile (small red point), and statistics
for matrices ordered by area, isolation,
and altitude, and for the matrix with
minimal nestedness temperature (Packed).
P = probability that a matrix is more
nested than a randomly ordered matrix

a potentially useful proxy for local biodiversity. At the same time, extinction. Most Nothobranchius species were endemic to one or
species turnover leads to variation in assemblage composition rather two adjacent regions, with a single species, N. melanospilus, found
than richness (Hill et al., 2017) across a particular set of assemblages. across all three major regions.
This highlights the need to conserve larger areas of potentially Although wetlands are now widely recognized as important
threatened habitats with a high contribution of species turnover to and valuable ecosystems, they remain largely unprotected and rap-
beta diversity. idly disappear in Africa and elsewhere (Dalu & Wasserman, 2022;
Regional pools of Nothobranchius species are formed through Gardner & Finlayson, 2018). The loss of natural habitat due to agri-
a combination of ecological and evolutionary processes. The in- culture expansion is the greatest threat to wetland fishes in Africa.
sular nature of seasonal wetland pools supports their isolation It contributes to 93% of all species loss followed by natural system
(Williams, 2006), including the evolution of many species with modification (38%), resident and commercial development (37%),
small, fragmented geographical ranges (van der Merwe et al., and pollution (27%; Nagy & Watters, 2021). We have collected mul-
2021). These range-­restricted species are intrinsically the most tiple species in highly modified habitats, including culverts along
endangered (Nagy & Watters, 2021). We have identified the re- the roads and active rice paddies. While this highlights the relative
gion of highest conservation priority (Mbezi region), where local perseverance of Nothobranchius populations and communities fac-
endemism is highest and beta diversity is driven by species turn- ing human pressure, it also illustrates the immediate threat to their
over. In the two largest regions (the basins of the lower Rufiji and habitats. Given their bright colors, extensive interest in their collec-
lower Ruvu rivers), we found that species richness was negatively tion by hobbyist breeders (Wildekamp, 2004), and their increasing
associated with altitude and that Nothobranchius assemblages use in biomedical and ecological research (Cellerino et al., 2016),
were nested due to selective colonization rather than selective Nothobranchius could serve as a flagship group for conservation
10 of 12 | REICHARD et al.

F I G U R E 5 Beta diversity (a) and its components (b: nestedness; c: species turnover) in three study area regions, calculated according to
Baselga (2010). Vertical lines represent range of parameters in 1000 random subsets of species-­richer sites, corresponding to the size of the
smallest site

efforts related to endangered small wetlands. Species-­specific male –­ review & editing (equal). Okinyi David Alila: Investigation (equal);
coloration enables easy and reliable visual identification to species Project administration (supporting); Writing –­ review & editing
and creates an opportunity for citizen science-­based conservation (equal). Matej Polačik: Conceptualization (supporting); Data cura-
efforts. tion (supporting); Investigation (equal); Methodology (supporting);
Writing –­ review & editing (equal).
AU T H O R C O N T R I B U T I O N
Martin Reichard: Conceptualization (lead); Data curation (sup- AC K N OW L E D G E M E N T S
porting); Formal analysis (supporting); Funding acquisition (lead); We thank Hieromin Lamtane and Nazael Madalla for facilitating
Investigation (equal); Methodology (lead); Project administration research and permits administration, two anonymous referees for
(lead); Resources (lead); Supervision (lead); Visualization (sup- helpful suggestions, and Rowena Spence for valuable comments
porting); Writing –­ original draft (lead); Writing –­ review & edit- on the manuscript and English correction. Funding came from the
ing (equal). Michal Janáč: Conceptualization (supporting); Data Czech Science Foundation (19-­01781S) to M.R.
curation (equal); Formal analysis (lead); Investigation (support-
ing); Methodology (equal); Resources (equal); Visualization (lead); C O N FL I C T O F I N T E R E S T
Writing –­ original draft (supporting); Writing –­ review & editing The authors declare no competing financial interests.
(equal). Radim Blazek: Conceptualization (supporting); Data cura-
tion (supporting); Investigation (equal); Validation (supporting); DATA AVA I L A B I L I T Y S TAT E M E N T
Visualization (supporting); Writing –­ review & editing (equal). Jakub Primary data are stored on Dryad (https://doi.org/10.5061/dryad.
Zak: Conceptualization (supporting); Investigation (equal); Writing m63xs​j44f).
REICHARD et al. | 11 of 12

García, D., Loureiro, M., Machín, E., & Reichard, M. (2019). Species co-­
ORCID occurrence and population dynamics in annual fish assemblages in
Martin Reichard https://orcid.org/0000-0002-9306-0074 the lower Río Uruguay basin. Environmental Biology of Fishes, 102,
Michal Janáč https://orcid.org/0000-0001-6264-6091 569–­580. https://doi.org/10.1007/s1064​1-­019-­0 0854​-­x
García, G., Loureiro, M., Berois, N., Arezo, M. J., Casanova, G., Clivio, G.,
Radim Blažek https://orcid.org/0000-0003-1150-0273
& Olivera, A. (2009). Pattern of differentiation in the annual killi-
Jakub Žák https://orcid.org/0000-0003-2845-8323
fish genus Austrolebias (Cyprinodontiformes: Rivulidae) from a bio-
Matej Polačik https://orcid.org/0000-0003-1309-2146 sphere reserve site in South America: A multidisciplinary approach.
Biological Journal of the Linnean Society, 98, 620–­635. https://doi.
REFERENCES org/10.1111/j.1095-­8312.2009.01303.x
García-­Navas, V., Martínez-­Núñez, C., Tarifa, R., Molina-­Pardo, J. L.,
Almeida-­Neto, M., Guimaraes, P., Guimaraes, P. R. Jr., Loyola, R. D., &
Valera, F., Salido, T., Camacho, F. M., & Rey, P. J. (2022). Partitioning
Ulrich, W. (2008). A consistent metric for nestedness analysis in eco-
beta diversity to untangle mechanisms underlying the assembly
logical systems: Reconciling concept and measurement. Oikos, 117,
of bird communities in Mediterranean olive groves. Diversity and
1227–­1239. https://doi.org/10.1111/j.0030-­1299.2008.16644.x
Distributions, 28(1), 112–­127. https://doi.org/10.1111/ddi.13445
Atmar, W., & Patterson, B. D. (1993). The measure of order and disorder
Gardner, R. C., & Finlayson, M. (2018). Global wetland outlook: State of the
in the distribution of species in fragmented habitat. Oecologia, 96,
world's wetlands and their services to people 2018. Secretariat of the
373–­382. https://doi.org/10.1007/BF003​17508
Ramsar Convention.
Bartáková, V., Nagy, B., Polačik, M., Blažek, R., Lamtane, H., & Reichard,
Gibson, L. A., Cowan, M. A., Lyons, M. N., Palmer, R., Pearson, D. J., &
M. (2020). Genetic diversity of a widespread annual killifish from
Doughty, P. (2017). Island refuges: Conservation significance of
coastal Tanzania. BMC Evolutionary Biology, 20(1), 1–­13. https://doi.
the biodiversity patterns resulting from ‘natural’fragmentation.
org/10.1186/s1286​2-­019-­1549-­2
Biological Conservation, 212, 349–­356. https://doi.org/10.1016/j.
Bartáková, V., Reichard, M., Blažek, R., Polačik, M., & Bryja, J. (2015).
biocon.2017.06.010
Terrestrial fishes: Rivers are barriers to gene flow in annual fishes
Helmstetter, A. J., Papadopulos, A. S., Igea, J., & Van Dooren, T. J. (2020).
from the African savanna. Journal of Biogeography, 42, 1832–­1844.
Trait evolution and historical biogeography shape assemblages of
https://doi.org/10.1111/jbi.12567
annual killifish. Journal of Biogeography, 47, 1955–­1965. https://doi.
Baselga, A. (2010). Partitioning the turnover and nestedness compo-
org/10.1111/jbi.13912
nents of beta diversity. Global Ecology and Biogeography, 19(1), 134–­
Hill, M. J., Heino, J., Thornhill, I., Ryves, D. B., & Wood, P. J. (2017). Effects
143. https://doi.org/10.1111/j.1466-­8238.2009.00490.x
of dispersal mode on the environmental and spatial correlates of
Bevilacqua, S., & Terlizzi, A. (2020). Nestedness and turnover unveil in-
nestedness and species turnover in pond communities. Oikos, 126,
verse spatial patterns of compositional and functional β-­diversity
1575–­1585. https://doi.org/10.1111/oik.04266
at varying depth in marine benthos. Diversity and Distributions, 26,
Lambert, J. W., Reichard, M., & Pincheira-­Donoso, D. (2019). Live fast,
743–­757. https://doi.org/10.1111/ddi.13025
diversify non-­adaptively: Evolutionary diversification of exception-
Brock, M. A., Nielsen, D. L., Shiel, R. J., Green, J. D., & Langley, J. D. (2003).
ally short-­lived annual killifishes. BMC Evolutionary Biology, 19(1),
Drought and aquatic community resilience: the role of eggs and
10. https://doi.org/10.1186/s1286​2-­019-­1344-­0
seeds in sediments of temporary wetlands. Freshwater Biology, 48,
Lévêque, C. (1997). Biodiversity dynamics and conservation: The freshwater
1207–­1218. https://doi.org/10.1046/j.1365-­2427.2003.01083.x
fish of tropical Africa. Cambridge University Press.
Brualdi, R. A., & Sanderson, J. G. (1999). Nested species subsets, gaps,
Loiseau, N., Mouquet, N., Casajus, N., Grenié, M., Guéguen, M., Maitner,
and discrepancy. Oecologia, 119, 256–­264. https://doi.org/10.1007/
B., Mouillot, D., Ostling, A., Renaud, J., Tucker, C., Velez, L.,
s0044​20050784
Thuiller, W., & Violle, C. (2020). Global distribution and conserva-
Cellerino, A., Valenzano, D. R., & Reichard, M. (2016). From the bush to
tion status of ecologically rare mammal and bird species. Nature
the bench: The annual Nothobranchius fishes as a new model sys-
Communications, 11(1), 1–­11. https://doi.org/10.1038/s4146​7-­020-­
tem in biology. Biological Reviews, 91, 511–­533.
18779​-­w
Costa, W. J. E. M. (2018). Comparative morphology, phylogeny and clas-
Loureiro, M., Borthagaray, A., Hernández, D., Duarte, A., Pinelli, V., &
sification of African seasonal killifishes of the tribe Nothobranchiini
Arim, M. (2015). Austrolebias in space: Scaling from ponds to bio-
(Cyprinodontiformes: Aplocheilidae). Zoological Journal of the
geographical regions. In N. Berois, G. García, & R. de Sá (Eds.),
Linnean Society, 184(1), 115–­135. https://doi.org/10.1093/zooli​
Annual fishes: Life history strategy, diversity, and evolution (pp. 111–­
nnean/​zlx102
132). CRC Press.
Costa, W. J. E. M. (2019). Two new species of seasonal killifishes of
Magurran, A. E. (2003). Measuring biological diversity. John Wiley & Sons.
the Nothobranchius melanospilus species complex from the East
Markham, J. (2014). Rare species occupy uncommon niches. Scientific
Africa biodiversity hotspot (Cyprinodontiformes: Aplocheilidae).
Reports, 4(1), 1–­3.
Vertebrate Zoology, 69(1), 73–­82.
Nagy, B., & Kis, A. (2010). Variation in habitat characteristics and the
Dalu, T., & Wasserman, R. J. (2022). Fundamentals of tropical freshwater
occurrence of Nothobranchius species in seasonal biotopes of
wetlands: From ecology to conservation management. Elsevier.
Tanzania. JAKA, 43, 130–­144.
Dalu, T., Wasserman, R. J., & Dalu, M. T. (2017). Agricultural intensifi-
Nagy, B., & Watters, B. (2021). A review of the conservation status of
cation and drought frequency increases may have landscape-­level
seasonal Nothobranchius fishes (Teleostei: Cyprinodontiformes),
consequences for ephemeral ecosystems. Global Change Biology,
a genus with a high level of threat, inhabiting ephemeral wetland
23, 983–­985. https://doi.org/10.1111/gcb.13549
habitats in Africa. Aquatic Conservation: Marine and Freshwater
Dorado, J., Vázquez, D. P., Stevani, E. L., & Chacoff, N. P. (2011). Rareness
Ecosystems, 32(1), 199–­216.
and specialization in plant-­pollinator networks. Ecology, 92(1), 19–­
Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., O’Hara,
25. https://doi.org/10.1890/10-­0794.1
R. B., Simpson, G. L., Stevenes, M. H. H., & Wagner, H. H. (2013).
Freiry, R. F., Gouvea, A., Becker, J., Lansac-­Tôha, F. A., Lansac-­Tôha, F.
Package ‘vegan’. Community ecology package, 2, 1-­295.
M., Pires, M. M., Stenert, C., & Maltchik, L. (2020). Community
Patterson, B. D., & Atmar, W. (1986). Nested subsets and the structure
structure and concordance patterns among zooplankton life stages
of insular mammalian faunas and archipelagos. Biological Journal
in subtropical temporary ponds. Aquatic Ecology, 54(1), 257–­270.
of the Linnean Society, 28(1–­2), 65–­82. https://doi.org/10.1111/
https://doi.org/10.1007/s1045​2-­019-­09740​-­1
j.1095-­8312.1986.tb017​49.x
12 of 12 | REICHARD et al.

Pepke, M. L., Irestedt, M., Fjeldså, J., Rahbek, C., & Jønsson, K. A. van der Merwe, P. D. W., Cotterill, F. P. D., Kandziora, M., Watters, B. R.,
(2019). Reconciling supertramps, great speciators and relict spe- Nagy, B., Genade, T., Flügel, T. J., Svendsen, D. S., & Bellstedt, D.
cies with the taxon cycle stages of a large island radiation (Aves: U. (2021). Genomic fingerprints of Palaeogeographic history: The
Campephagidae). Journal of Biogeography, 46, 1214–­1225. https:// tempo and mode of Rift tectonics across tropical Africa has shaped
doi.org/10.1111/jbi.13577 the diversification of the killifish genus Nothobranchius (Teleostei:
Polačik, M., Harrod, C., Blažek, R., & Reichard, M. (2014). Trophic niche Cyprinodontiformes). Molecular Phylogenetics and Evolution, 158,
partitioning in communities of African annual fish: Evidence 106988. https://doi.org/10.1016/j.ympev.2020.106988
from stable isotopes. Hydrobiologia, 721, 99–­106. https://doi. van der Valk, A. (2012). The biology of freshwater wetlands. Oxford
org/10.1007/s1075​0 -­013-­1652-­0 University Press.
Polačik, M., & Reichard, M. (2010). Diet overlap among three sym- Vanschoenwinkel, B., Waterkeyn, A., Nhiwatiwa, T., Pinceel, T.,
patric African annual killifish species Nothobranchius spp. from Spooren, E., Geerts, A., Clegg, B., & Brendonck, L. (2011).
Mozambique. Journal of Fish Biology, 77, 754–­768. https://doi. Passive external transport of freshwater invertebrates by el-
org/10.1111/j.1095-­8649.2010.02717.x ephant and other mud-­wallowing mammals in an African sa-
Polačik, M., Vrtílek, M., Reichard, M., Žák, J., Blažek, R., & Podrabsky, J. vannah habitat. Freshwater Biology, 56, 1606–­1619. https://doi.
(2021). Embryo ecology: Developmental synchrony and asynchrony org/10.1111/j.1365-­2427.2011.02600.x
in the embryonic development of wild annual fish populations. Vrtílek, M., Žák, J., Pšenička, M., & Reichard, M. (2018). Extremely
Ecology and Evolution, 11, 4945–­4956. https://doi.org/10.1002/ rapid maturation of a wild African annual fish. Current Biology, 28,
ece3.7402 R822–­R824. https://doi.org/10.1016/j.cub.2018.06.031
R Core Team. (2018). R: A language and environment for statistical comput- Watters, B. R., Nagy, B., van der Merwe, P. D., Cotterill, F. P., & Bellstedt,
ing. R Foundation for Statistical Computing. D. U. (2020). Redescription of the seasonal killifish species
Reichard, M. (2015). Annual fishes: Life history strategy, diversity, and Nothobranchius ocellatus and description of a related new species
evolution. In N. Berois, G. Garcia, & R. O. De Sa (Eds.), The evolution- Nothobranchius matanduensis, from eastern Tanzania. Ichthyological
ary ecology of African annual fishes (pp. 133–­158). CRC Press, Taylor Exploration of Freshwaters, 30, 151–­178.
and Francis group. Wildekamp, R. (2004). A world of killies (vol. 4). Atlas of the Oviparous
Reichard, M., Janáč, M., Polačik, M., Blažek, R., & Vrtílek, M. (2017). Cyprinodontiform fishes of the world. NHBS Academic & Professional
Community assembly in Nothobranchius annual fishes: Nested pat- Books.
terns, environmental niche and biogeographic history. Ecology and Wildekamp, R. W. (2019). Back to the theme Nothobranchius guentheri
Evolution, 7, 2294–­2306. and Nothobranchius melanospilus, with additional notes to the N.
Reichard, M., & Polačik, M. (2019). The Natural History of Model melanospilus complex sensu Costa 2019. JAKA, 52, 43–­52.
Organisms: Nothobranchius Furzeri, an ‘instant’ Fish from an Williams, D. D. (2006). The biology of temporary waters. Oxford University
Ephemeral Habitat. eLife, 8, e41548. https://doi.org/10.7554/ Press.
eLife.41548 Wright, D. H., Patterson, B. D., Mikkelson, G. M., Cutler, A., & Atmar, W.
Reichard, M., Polačik, M., Blažek, R., & Vrtílek, M. (2014). Female bias in (1997). A comparative analysis of nested subset patterns of spe-
the adult sex ratio of African annual fishes: Interspecific differences, cies composition. Oecologia, 113(1), 1–­20. https://doi.org/10.1007/
seasonal trends and environmental predictors. Evolutionary Ecology, s0044​20050348
28, 1105–­1120. https://doi.org/10.1007/s1068​2-­014-­9732-­9 Xu, A., Han, X., Zhang, X., Millien, V., & Wang, Y. (2017). Nestedness of
Roberts, T. R. (1975). Geographical distribution of African freshwater butterfly assemblages in the Zhoushan Archipelago, China: Area ef-
fishes. Zoological Journal of the Linnean Society, 57, 249–­319. https:// fects, life-­history traits and conservation implications. Biodiversity
doi.org/10.1111/j.1096-­3642.1975.tb018​93.x and Conservation, 26, 1375–­1392. https://doi.org/10.1007/s1053​
Shidlovsky, K. (2010). Collecting Nothobranchius in south-­western 1-­017-­1305-­0
Tanzania. Nothobranchius Archives, 1, 18–­27.
Socolar, J. B., Gilroy, J. J., Kunin, W. E., & Edwards, D. P. (2016). How
should beta-­diversity inform biodiversity conservation? Trends
How to cite this article: Reichard, M., Janáč, M., Blažek, R.,
in Ecology & Evolution, 31(1), 67–­8 0. https://doi.org/10.1016/j.
tree.2015.11.005
Žák, J., Alila, O. D., & Polačik, M. (2022). Patterns and drivers
Soininen, J., Heino, J., & Wang, J. (2018). A meta-­analysis of nested- of Nothobranchius killifish diversity in lowland Tanzania.
ness and turnover components of beta diversity across organisms Ecology and Evolution, 12, e8990. https://doi.org/10.1002/
and ecosystems. Global Ecology and Biogeography, 27(1), 96–­109. ece3.8990
https://doi.org/10.1111/geb.12660
Ulrich, W. (2009). Nestedness analysis as a tool to identify ecological gra-
dients. Ecological Questions, 11, 27–­3 4. https://doi.org/10.12775/​
v1009​0 -­0 09-­0 015-­y