Source data

The dataset was obtained in two parts, from the online New Zealand Organisms Register (NZOR) (Anonymous 2018) and an original digital version for animals only whose original source was the three volumes edited by Gordon (2009, 2010, 2012). It includes not only described species, but candidate species yet to be described and which occur in specimen collections. These undescribed taxa are assumed to be likely to be described as new species and thus counted here as equal to named species so the dataset is a more accurate indicator of the total biotic diversity. These sources were compared with each other and the summary published by Gordon (2013). Differences between datasets arose due to alternative taxonomic classifications, including different hierarchical arrangements of phyla and classes; e.g. Charophyta within Chlorophyta in NZOR. Following Gordon (2013), Anthocerotophyta and Marchantiophyta are included with Bryophyta, and Dinophyta (Dinoflagellata) with Myzozoa (Miozoa), and Radiozoa with Foraminifera. Minor errors in attribution of species as living and fossil (e.g. most Hemichordata in NZOR were fossil species), and by marine, terrestrial and freshwater environment (e.g. Pycnogonida mistakenly included as terrestrial), were corrected. Instances of a species being recorded twice with different spellings of authority names occurred, and instances of higher taxa being included in the species names field were removed prior to analysis. Curiously, the only Phylum that was in the source volumes but not available in the digital data was that comprising the three native (not endemic) species of penis worms, Priapulida. The final data used is available as Supplementary Material.

Marine species were recorded for the New Zealand Exclusive Economic Zone and thus range from sub-tropical latitudes to sub-Antarctic islands, and the intertidal to deep-sea. Only taxa classified by endemicity and environment were included in this analysis. While the numbers of exotic species are shown per taxon, they are not included in the analyses.

The number of species and percent endemicity were provided at different levels in the taxonomic hierarchy as the available data permitted (). Phyla with less than five species were not tabulated, namely: Dicyemida (6 marine endemic species), Echiura (7 spp), Gastrotricha (1 freshwater, 4 marine), Nematomorpha (3 freshwater and 1 marine endemic, 1 native freshwater), Phoronida (3), Loricifera (4 marine), Orthonectida (1 marine), Archaea (5), Glaucophyta (1 freshwater). Neither were the Charophyta due to uncertainty in their classification (29 spp in NZOR included in Chlorophyta but 527 reported in Gordon (2013) with 43 endemic). The total fungi present is similar in NZOR and Gordon (2013), but richness of Oomycota (Pseudofungi) varied (307 vs 161) and the proportions in each environment were not captured sufficiently in NZOR for environmental comparisons.

The number of accepted species per taxon was obtained from the World Register of Marine Species (WoRMS) for marine and some associated taxa (Costello et al. 2013a, Horton et al. 2022). Several taxa in WoRMS include all their freshwater and marine relatives, including the crustaceans and molluscs. These data were cross-checked against recorded and estimated numbers of species in taxa in the Catalogue of Life (Bánki et al. 2022), including for freshwater and terrestrial species. Where differences in the taxonomic classifications between these three sources made comparisons questionable these data were excluded from the analyses. With two million species listed, the CoL is an almost complete checklist of all accepted species (Costello et al. 2022). Thus, the recorded number of species in most taxa approximates the number of accepted species, notwithstanding that new species are still being named, names are being synonymised and classifications revised (Costello et al. 2013a, Costello 2020).

Biological traits

In the dataset, all birds and insects were classified as terrestrial, even though some birds are closely associated with freshwater and/or marine environments, and several orders of insects live most of their lives in freshwaters. Because only 2% of the Plantae species were classified by environment, comparisons across environments were dominated by animals, microbes (i.e. Bacteria, Archaea, Protozoa) and Chromista. Taxa were attributed a typical body size as microscopic if 1 mm or less, megafauna if typically > 10 cm and macroscopic in-between these values, following Costello et al. (2013b). As many marine taxa have both pelagic and benthic life-stages, and data on their relative duration and significance are not available at species level, those that were almost entirely pelagic and benthic were compared: pelagic being Chaetognatha, Ctenophora, Cubozoa, Mammalia, Reptilia, Scyphozoa; and benthic the peracarid crustaceans which lack planktonic larvae, namely Amphipoda (excluding 101 species in New Zealand associated with pelagic hosts), Cumacea, Isopoda and Mysida (). All birds and insects were classified as ‘flighted’ because it is assumed their first colonists of New Zealand could fly, although this ability varies today. The Phthirptera were included as flighted insects because their hosts, birds, fly. However, the also flightless flea parasites, Siphonaptera, were not classified as flighted because about one third of species in New Zealand parasitise mammals (excluding bats) so the estimate of flighted species is conservative. The classification kept the Entognatha, namely the wingless arthropods, Collembola, Diplura, Protura and Thysanura separate from the Insecta. Most vertebrates, arthropods and annelids were classified as mobile, although there are exceptions in most higher taxa ().

Richness

Of the total 54,000 species, most were animals (63%), invertebrates (60%) and arthropods (38%) (). Molluscs and chordates (mostly fish) comprised 9% and 4% respectively. There were similar proportions of Plantae, fungi and Chromista (14%, 13%, 8%) and few prokaryotes (Bacteria, Archaea, Protozoa, 2%) ().

The number of species per taxon in New Zealand was highly correlated with the number globally for both marine, and for terrestrial and freshwater combined (and ). However, there are differences. Excluding fungi, because 74% are introduced species but were not classified as such in the present database, the percentages for the world and New Zealand are: 78% and 65% of species are animals, 64% and 38% arthropods, 54% and 20% insects, 19% and 14% plantae and reversed with 1% and 8% Chromista. The relatively higher proportion of Chromista may be due to marine taxa comprising 30% of the New Zealand biota but 12% of the global biota.

Figure 2. (A) The total number of marine species recorded in New Zealand was highly correlated (R² 0.92) with the number known in the world for the higher taxa (). Even when the three most species rich taxa are excluded the correlation was still high (R² 0.86). (B) The relationship between the number of marine species globally (red X symbols) and in New Zealand (hollow circles symbols) to percent endemicity of the higher taxa (excluding samples with less than ten species). Neither of these variables was well correlated with the percent endemicity in New Zealand (R² 0.05) or globally (R² 0.07).

Figure 3. The total number of freshwater and terrestrial species recorded in New Zealand is highly correlated with the number known in the world (upper plot, R²0.95). When the three most species rich taxa are excluded the correlation is still strong (R² 0.79). However, neither of these variables is well correlated with the percent endemicity in New Zealand (red X symbol) and globally (hollow circle symbol) (both R² 0.01) (lower plot, excluding samples with less than ten species).

Comparing across environments, the terrestrial harboured 65%, marine 30% and freshwater 5% of all species. However, while most and the greatest proportion of introduced species were terrestrial at 32%, fewest were marine at 1% ().

Endemicity

There was no correlation between percent endemicity and total species richness per taxon in New Zealand, globally, nor in marine and terrestrial environments (and ). Thus, percent endemicity per taxonomic (i.e. phylogenetic) group does not help predict species richness at regional or global scales.

Generally, terrestrial taxa had higher endemicity and marine the lowest (see mean and quartile distributions in ). Within each environment there were taxa with very low and very high endemicity (). Overall, 52% of native species were endemic, with 23%, 39% and 68% for freshwater, marine and terrestrial environments respectively (). The average percent endemicities across taxa were 44%, 36% and 65% per environment respectively ().

Figure 4. The distribution of percent endemicity across freshwater, marine and terrestrial taxa recorded in New Zealand, showing the mean (x), median (horizontal bar), range and upper and lower quartiles (box), where each data point is a taxon.

Dispersal traits

Microscopic and marine taxa tended to have lower levels of endemicity (). In addition, benthic marine taxa have greater endemicity than pelagic. A lower proportion of terrestrial taxa occur at lower levels of endemicity ().

Microscopic taxa and megafauna had the lowest endemicity, followed by sedentary taxa, especially for freshwater taxa (). Mobile, flighted and macro taxa had the highest endemicity across all environments, with 91-94% of these terrestrial trait groups endemic ().

Figure 5. The average percent endemicity of taxa grouped by traits across (A) freshwater (black bar), marine (hollow bar), terrestrial (grey bar) and all (hatched bar) taxa recorded in New Zealand as a bar chart and (B) table. Because the totals are based on the number of species they are not the same as the average of percentages for each environment.

There was exceptionally high endemicity of flighted freshwater and terrestrial taxa, suggesting poor dispersal by flight (, ). While endemicity was generally lower in the marine environment, it was four to five times higher for microscopic and sedentary animal taxa than in freshwater and terrestrial environments (). Also, in the marine environment, the 97 species of entirely pelagic and 803 species of entirely benthic taxa had 9% and 75% endemicity respectively. While there was no significant correlation between percent endemicity of taxa and their regional and global species, there was when taxa were grouped by these traits ().

Figure 6. The relationship between percent endemicity to the number of species in each of the trait groups by environment in B table of body size, pelagic-benthic, mobile-sedentary and flight (n = 18, red dots, R² 0.37). Hollow circles are the totals across the environments and excluded from the correlation to avoid double counting.

Environment

The results confirmed the expected lower endemicity of marine compared to terrestrial species. This reflects the natural isolation of the land by the ocean while an artificial boundary, the EEZ, is used to bound the marine biota. Furthermore, globally, marine species endemicity decreases with distance from the coast and depth as the environment becomes more stable and homogenous (Costello et al. 2017, 2018; Sayre et al. 2017). Thus, marine species generally inhabit wider depth and geographic ranges with depth, and marine species endemicity is greater in shallow depths.

Freshwater endemicity (excluding freshwater insects) was similar or lower to marine which contradicted expectations. Because most freshwater habitats are temporary on evolutionary timescales, their fauna and flora need to have effective dispersal, whether aerial, through coastal seas, or by attachment to animals (phoresy). For example, Dehling et al. (2010) found species associated with apparently more isolated standing waters (e.g. lakes, ponds) were more widely distributed than species in running waters. Thus, the lower endemicity of freshwater than terrestrial species in New Zealand seems due to the temporary nature of their habitats, and consequently their biota have good dispersal mechanisms.

Body size

The higher endemicity and species richness of benthic compared to pelagic, and macrofauna compared to microscopic and mega-fauna, were also confirmed. Thus, the largest and smallest species tend to be more cosmopolitan and thus likely to have higher gene flow and fewer species, explaining their previously reported species richness globally (Costello et al. 2013b). In the ocean, the smallest and largest pelagic taxa are the most widespread and have fewer species compared to macrofauna (Costello et al. 2017). However, 37% endemicity for terrestrial megafauna in New Zealand is not low and reflects long term isolation and distance from other continents.

Macrofauna

That macrofauna dominate species richness from local to global scales is clear (Costello et al. 2013b). The New Zealand data show this size group also has exceptionally high endemicity, 78% overall and 65%, 52% and 93% across freshwater marine and terrestrial environments respectively (). Yet, the fact that most arthropods are mobile, and generally most insects can fly, could lead to high dispersal and gene flow and low speciation rates. If this is true, then evidently other factors are driving diversification (speciation minus extinctions) apart from potential mobility. The ability of mobile macrofauna to specialise in geographically scattered microhabitats and food sources, including becoming parasitic, opens opportunities for ecological niche differentiation (Costello and Chaudhary 2017). However, they are at high risk of predation by highly mobile vertebrates. The need to avoid predation is apparent in the nocturnal emergence of benthic freshwater (so called ‘freshwater drift’ in rivers), marine (including benthos emergence and vertical migration of pelagic species) and terrestrial arthropods (McLeod and Costello 2017). Thus, in addition to niche specialisation macrofauna are under continuous pressure to reduce predation risk, which is likely a stronger driver of selection than food supply because death limits evolution more than going hungry (the ‘life-dinner’ hypothesis, Dell et al. 2011). Thus, the high endemicity of mobile macrofauna may be a consequence of opportunities for niche specialisation and the need to limit predation.

Preliminary support for a role of predation is that marine macro-epi-benthic diversity is lower in the tropics than polar latitudes, and inversely correlated with fish species’ abundance (Edgar et al. 2017). Furthermore, the diversity of amphipod crustaceans seems surprisingly low in the Coral Triangle, considered the most rich marine region for fish and corals, compared to higher latitudes (Arfianti and Costello 2021). There is growing experimental evidence of higher tropical predation rates in marine pelagic and benthic environments (Rodemann and Brandl 2017; Roesti et al. 2020; Whalen et al. 2020; Freestone et al. 2021; Ashton et al. 2022), and this seems likely to be the case in terrestrial and freshwater environments for the same reasons despite the variable evidence in the literature (Moles and Ollerton 2016). However, predation is only one selection pressure, and the continuously warm tropical climate provides better growing conditions than in the high latitudes where winter dampens growth, and larger populations and shorter generation times provide more opportunities for speciation regardless of predation pressure (reviewed in Costello and Chaudhary 2017).

Flight

The high endemicity of flighted taxa was not expected, as it may be assumed that species that can fly would be widely dispersed. However, this may only hold across continents and nearby islands for terrestrial and freshwater species. New Zealand appears to be too far from other landmasses for regular colonisation by most flying species. Indeed, New Zealand exemplifies the well-known phenomenon of a loss of flight in island insects and terrestrial birds which further restricts their dispersal (e.g. Darwin 1860; Diamond 1984; Costello 1988; Veale et al. 2018; Foster et al. 2021) and geographic range (McCulloch et al. 2017).

Richness

The number of species per taxon in New Zealand was highly correlated with the number named globally. Positive correlations have similarly been observed between the number of species of trees in the same taxa between continents (Silva de Miranda et al. 2022), and of new species being named across all taxa and the number of named species (Costello et al. 2013b). However, neither New Zealand nor global richness of taxa correlated with percent endemicity of the New Zealand biota. This contradicted expectations because it was hypothesised that if a taxon had high endemicity then it should have low dispersal and geneflow, leading to high rates of spatial speciation, and high global species richness. In contrast, when taxa were grouped by the traits of body size and mobility there was a significant relationship between endemicity and richness (). These contrasting relationships suggest that the dispersal traits of a taxon are more important in leading to new species evolving (speciation) that its phylogeny. Indeed, major marine and terrestrial lineages have gone extinct in the past, significantly reducing their present species richness, and other taxa diversified (Mayhew et al. 2008; Yasuhara et al. 2020). Nevertheless, the variability in the endemicity-richness relationships indicates that either the underlying data are at too coarse a taxonomic resolution, are not sufficiently complete, and/or that other traits and factors are also important in addition to dispersal.