Diversity of Spring Invertebrates and Their Habitats: A Story of Preferences

Springs, as unique ecotonal habitats between surface and hypogean areas, are considered endangered aquatic ecosystems due to direct and indirect human impacts and climate change issues. They are distinctive water habitats that are often inhabited by a diverse but mostly stenotypic group of organisms. The present study considered 31 springs from the Apuseni Mountains (the Romanian Carpathians) that were classified as rheocrene, helocrene, and limnocrene based on their geomorphology and hydrology. Samples from three substrate types (rocks, sand, and bryophytes) were collected using standard methods for crenic invertebrates. A total of 64,462 individuals belonging to 17 invertebrate taxa were identified: aquatic worms, mollusks, crustaceans, water mites, and insects. Amphipoda and Diptera–Chironomidae were the dominant taxa in most springs. At a community level, patterns of habitat preference were demonstrated for 12 invertebrate groups using the standardized selection index (B) and expressed as the number of springs where a certain group selected rocks, sand, and/or bryophytes: Four groups exhibited preferences for bryophytes (Coleoptera, Diptera, Plecoptera, and Trichoptera), Ephemeroptera exhibited preferences for rocks, and Copepoda exhibited preferences for sand. Amphipoda, Platyhelminthes, and Ostracoda displayed preferences for all three substratum types, while Gastropoda, Hydrachnidia, and Oligochaeta recorded lower percentages in springs where habitat preferences were significant. In addition, crenic invertebrates were divided into three guilds, depending on their dispersion abilities in any stage of their life cycle: sedentary (not-winged groups), mobile (winged groups), and ectoparasites (water mites that were able to leave the springs on their winged hosts). Sedentary taxa recorded higher percentages of abundances and habitat preferences towards rocks and sand, while ectoparasites (Hydrachnidia) and the mobile guilds tended to prefer bryophytes. This segregation might be explained by individual adaptations to the particularities of each type of substratum, such as the bodily form of the copepods, which are well suited for sand interstices, a habitat that our data showed that they preferred. Our results represent novel contributions to the knowledge of habitat preferences of spring invertebrates from the Apuseni Mountains, adding value to similar data from the Western Carpathians, the Alps, and the Dinaric region.


Introduction
Springs, which are defined as the place of appearance on the surface of groundwaters as a result of sediment permeability [1], represent unique ecotonal ecosystems where groundwater, surface waters, and terrestrial ecosystems interact to form environments characterized by a high biological biodiversity, relatively stable environmental conditions, and, usually, good water quality [1][2][3][4]. The main drainage basins from the study area belong to the Somes , ul Cald, Aries , , and Cris , uri Rivers. The vast majority of the sampling springs are permanent with annual discharge variations, but without complete drought phases [49]. One intermittent spring was considered, with regular fluctuations in discharge (RI-rheocrene intermittent, Table 1). Two sampling springs were characterized by standing water (L-limnocrene), while the other two originated from one or several point sources, but formed a swampy slow-flowing area (H-helocrene). Most of the 31 sampling springs were rheocrene (R), with a point source and a fast-flowing course that either came out of a cave (RC-rheocrene cave) or formed a marsh with a higher water current (RH-rheohelocrene). All spring type categories followed Cantonati et al. (2012) and Zollhofer et al. (2000) [1,50].

Sampling Methods
Physical and chemical parameters were measured in the field; the water temperature, pH, TDS, and conductivity were recorded using the Hanna HI98130 and Hanna HI98194 portable multimeters, while dissolved oxygen concentrations were measured using the YSI-52 oximeter and Hanna HI98194 multimeter. The relative surfaces of three major habitats (rocks, sand, and bryophytes) were estimated in situ and recorded as percentages.
Invertebrate samples were collected using benthic nets (250 µ m mesh size) from the spring eucrenon area. Five sampling points and a standardized sampling time of approximately 5 min were considered for each main substratum type (rocks/sand/bryophytes). Mineral sediments not exceeding 0.2 cm were considered as "sand", while gravel, pebbles, and stones (size greater than 0.2 cm, up to 20 cm) were assessed as rocks, following the AQEM consortium's protocol [51]. Only submerged habitats were taken into consideration when estimating the relative surface of every substrate type in the moment of the sampling. Rocky and sandy samples were collected by disturbing the sediments and placing the net downstream from the area of interest, while for the bryophyte samples, stirring

Sampling Methods
Physical and chemical parameters were measured in the field; the water temperature, pH, TDS, and conductivity were recorded using the Hanna HI98130 and Hanna HI98194 portable multimeters, while dissolved oxygen concentrations were measured using the YSI-52 oximeter and Hanna HI98194 multimeter. The relative surfaces of three major habitats (rocks, sand, and bryophytes) were estimated in situ and recorded as percentages.
Invertebrate samples were collected using benthic nets (250 µm mesh size) from the spring eucrenon area. Five sampling points and a standardized sampling time of approximately 5 min were considered for each main substratum type (rocks/sand/bryophytes). Mineral sediments not exceeding 0.2 cm were considered as "sand", while gravel, pebbles, and stones (size greater than 0.2 cm, up to 20 cm) were assessed as rocks, following the AQEM consortium's protocol [51]. Only submerged habitats were taken into consideration when estimating the relative surface of every substrate type in the moment of the sampling. Rocky and sandy samples were collected by disturbing the sediments and placing the net downstream from the area of interest, while for the bryophyte samples, stirring and washing techniques were deployed [52]. Three final samples, one for each substrate type, were collected from each spring. Samples were preserved in the field in 96% ethyl alcohol. In the laboratory, they were sorted and analyzed under Nikon SMZ 645 and 800 stereomicroscopes and Nikon YS100 microscopes. Crenic invertebrates were identified at various taxonomic levels (species for copepods, genera for water mites, families for Diptera, higher taxa for other groups) using standard keys for aquatic invertebrates [53][54][55][56][57][58][59][60].
Crenic invertebrates were divided into three ecological guilds depending on their ability to disperse: sedentary groups, ectoparasites, and mobile forms. Sedentary forms included non-emergent groups-regardless of body size-that spent all stages of their life cycles in spring habitats, without a winged phase [27]. Ten groups were characterized as sedentary in the 31 sampling sites: Oligochaeta, Nematoda, Platyhelminthes, Hirudinea, Gastropoda, Bivalvia, Amphipoda, Isopoda, Ostracoda, and Copepoda. In contrast, mobile groups were emergent taxa, which were characterized by winged adults. As a consequence, they had a better dispersion ability. In our dataset, mobile forms included the insect groups of Coleoptera, Diptera (Chironomidae and other families), Ephemeroptera, Heteroptera, Plecoptera, and Trichoptera. The particular life cycle of Hydrachnidia placed them in a third dispersion guild: ectoparasites. Water mites display a particularly complex change in morphology and feeding behavior during their development; larvae are obligate parasites of freshwater insects (often with high selectivity for hosts), and are thus able to leave the aquatic habitat. Active post-larval stages-deutonymphs and adults-feed as predators, mostly on eggs and on aquatic invertebrates, especially Diptera [61].

Statistical Methods
The relative percentages of abundance (percentage of each group) and frequency (percentage of springs where certain groups were present) were calculated for the crenic invertebrate communities. Diversity was expressed as the dominance index [62]: D = Σ(n i /n) 2 , where: D = dominance; n i = number of individuals from group i; n = total number of individuals. The dominance ranged from 0 (all groups were equally represented) to 1 (only one group dominated).
The selection index [63][64][65] was used to measure habitat preferences for rocks, sand, and bryophytes in each crenic invertebrate group from the 31 sampling springs. The standardized ratio was used, which summed to 1.00 for all resource types: B = (pu i /pm i )/Σ(pu i /pm i ), where: B = standardized selection index for group i; pu i = proportion of individuals in each group using resource i; pm i = proportion of resource i in the environment. B values higher than 0.33 (1/number of resources) indicated a preference. The G-test was used to test the null hypothesis of equal use of all three habitats (H0 = animals select resources at random). The chi-squared value (χ 2 ) was calculated and compared to the critical value of 5.99 (p = 0.05) or 9.21 (p = 0.01) with 2 degrees of freedom according to the recommendations made by Manley et al. (1993,2002) [63,64].
Principal component analysis (PCA) was used to observe aggregation trends in the data based on spring type and substratum category. Multivariate analyses were performed using the statistical package CANOCO version 5.15 with the unconstrained ordination menu [66].

Crenic Habitats
The three substrate categories considered for the present study (rocks, sand, and bryophytes) were present in all 31 sampling springs, regardless of their type (fast-flowing, lentic, or paludal) (Supplementary Table S1). A PCA biplot with supplementary variables was constructed in order to summarize the similarities in environmental variable values (substrate types) among sampling springs ( Figure 2). Strong positive correlations were recorded in the case of sand percentages and the horizontal axis, as well as in the case of rocky substrate percentages and the vertical axis, while a negative correlation was recorded between bryophyte percentages and the horizontal axis. The biplot shows no clear grouping of springs (rheocrene, limnocrene, or helocrene) according to the percentages of the three substrate types. rocky substrate percentages and the vertical axis, while a negative c orded between bryophyte percentages and the horizontal axis. The bi grouping of springs (rheocrene, limnocrene, or helocrene) according to the three substrate types.  Table 1.
Most sampling springs were rheocrene, while two were helocre two were limnocrene (4 and 19). The physico-chemical parameters showed overall homogeneous variations: water temperature was reco average (±1.83 °C standard deviation); the pH was around neutral (7.7 conductivity was recorded as having generally low values (336.13 ± 12 mentary Table S1). Dissolved oxygen (8.82 ± 2.14 mg/L) was recorde values in slow-flowing springs (limnocrene), as expected. The very from springs 1 and 7 were caused by reduced water levels and flowin by alterations within the spring bed.

Crenic Invertebrate Communities
A total of 64,462 crenic invertebrates were identified in the 31 spri diversity of taxa ( Figure 3). The total number of individuals collected springs amounted to an average value of 2079.42 ± 1730.71 (expressed a ard deviation), ranging from a minimum of 156 to a maximum of 847 phipoda and Chironomidae (Diptera) were the only groups with a freq they were present in all samples and in high numbers (Figure 3, supple Other constant groups with a frequency higher than 70% were the follo  Table 1. Most sampling springs were rheocrene, while two were helocrene (15 and 24) and two were limnocrene (4 and 19). The physico-chemical parameters measured in situ showed overall homogeneous variations: water temperature was recorded as 8.88 • C on average (±1.83 • C standard deviation); the pH was around neutral (7.76 ± 0.49); the water conductivity was recorded as having generally low values (336.13 ± 127.31 µS/cm, Supplementary Table S1). Dissolved oxygen (8.82 ± 2.14 mg/L) was recorded as having lower values in slow-flowing springs (limnocrene), as expected. The very low oxygen values from springs 1 and 7 were caused by reduced water levels and flowing speeds, as well as by alterations within the spring bed.

Crenic Invertebrate Communities
A total of 64,462 crenic invertebrates were identified in the 31 springs, belonging to a diversity of taxa ( Figure 3). The total number of individuals collected from the sampling springs amounted to an average value of 2079.42 ± 1730.71 (expressed as the mean ± standard deviation), ranging from a minimum of 156 to a maximum of 8475 individuals. Amphipoda and Chironomidae (Diptera) were the only groups with a frequency of 100%, i.e., they were present in all samples and in high numbers (Figure 3, Supplementary  Table S1). Other constant groups with a frequency higher than 70% were the following: Gastropoda, Platyhelminthes, Hydrachnidia, Coleoptera, Ephemeroptera, Plecoptera, Trichoptera, and other families of Diptera (considered as a separate group that included Anthomydae, Athericidae, Ceratopogonidae, Culicidae, Dixidae, Empididae, Limoniidae, Psychodidae, Ptychopteridae, Simuliidae, Stratiomyidae, Tabanidae, and Tipulidae). stratum. Only Ljania sp. was present in sandy habitats and not in bryop Eight copepod species (from a total of 275 individuals) were ident poid (Copepoda: Cyclopoida, three species) and harpacticoid (Copepod 5 species), adults (232), and immature copepodites (43). All species were gophyle forms, which were previously reported in groundwater habita   Most invertebrate communities (>65%) included evenly distributed taxa, recording dominance values below 0.40, while in a few cases, amphipods or chironomids exceeded 90% of the total number of individuals, causing the dominance values to go higher than 0.70.
Hydrachnidia and Copepoda were identified on the genus and species levels, respectively (Table 2), because they represent important, widely encountered components in all types of springs [65,67,68]. They are both meiofaunal groups that contribute to spring biodiversity [31,69]. A total of 329 water mites were collected, belonging to eight genera. More than 60% were deutonymphs, which were similar to the adults, but smaller and sexually immature. Four water mite taxa appeared in the bryophyte samples alone, while four genera had a higher frequencies and percentages of abundance in this type of substratum. Only Ljania sp. was present in sandy habitats and not in bryophytes ( Table 2).

Habitat Preferences at a Community Level
The habitat preferences in the case of crenic invertebrate communities, which were tested using the standardized selection index (B), varied extensively for the 17 taxonomic groups. The results were expressed as the percentage of springs (with respect to the total number of 31) where invertebrate groups recorded the highest frequency of significant B values (selected using the G test), i.e., the number of springs where a certain taxon or guild selected rocks, sand, and/or bryophytes. No significant preferences were observed for Nematoda, Hirudinea, Bivalvia, Isopoda, and Heteroptera (according to the G test; χ 2 , for p < 0.05) (Supplementary Table S2). Our results show that amphipods were found in high percentages in springs where they preferred either rocks, sand, or bryophytes (45%, 32%, and 41%, respectively) ( Figure 4). Similar findings concerned Platyhelminthes and Ostracoda, but with lower percentages (6%, 3%, and 6%, respectively, for Platyhelminthes and 12%, 9%, and 12%, respectively, for Ostracoda).

Habitat Preferences at a Guild Level
Three guilds were analyzed with respect to the number of individuals found in each sampling spring: sedentary forms, mobile forms, and ectoparasites. Transformed logarithmic values were used due to the high number of "0" values in the dataset ( Figure 5).
Sedentary forms recorded higher relative abundances compared to the mobile ones on rocky and sandy substrata (they reached higher abundances on rocks in 55% of springs and on sand in 65% of springs). Mobile groups had higher numbers in bryophytes in 77% of all sampling points.
The standardized index of preferences (B values > 0.33) depicted similar trends, with sedentary forms selecting rocky and sandy substrata in 22% and 64% of all springs, respectively. Mobile forms, however, showed preferences for sand and bryophytes in 64% A rocky substratum was selected by Ephemeroptera in the case of 25% of rocky habitat springs (Figure 4a). Sand was the preferred substratum for Copepoda (Supplementary  Table S2), while bryophytes were selected in a higher percentage by Coleoptera, Diptera (Chironomidae and other families), Plecoptera, and Trichoptera (Figure 4b,c; Supplementary  Table S2).
On the other hand, the other three groups, Gastropoda, Hydrachnidia, and Oligochaeta, recorded lower percentages in springs where their habitat preferences were significant. Hydrachnidia, for example, selected rocks, sand, and bryophytes in 3%, 0%, and 3% of all springs, respectively.

Habitat Preferences at a Guild Level
Three guilds were analyzed with respect to the number of individuals found in each sampling spring: sedentary forms, mobile forms, and ectoparasites. Transformed logarithmic values were used due to the high number of "0" values in the dataset ( Figure 5). and 51% of all springs, respectively. Ectoparasites were better represented in bryophyte substrata in terms of relative abundance and frequency ( Figure 5).

Figure 5.
Relative abundance of crenic invertebrate guilds, labeled as sedentary, ectoparasitic, and mobile on rocks (top), sand (middle), and bryophytes (down); the abundance values were transformed as follows: log(value + 1); numbers in orange squares: springs with a higher number of sedentary forms; numbers in green squares: springs with a higher number of mobile forms; sampling codes from 1 to 31, as in Table 1.

Habitat Preferences for Water Mites and Copepods
The habitat preferences for Hydrachnidia, which were calculated using the standardized selection index (B), showed that most water mite genera selected bryophytes, while Lebertia sp. and Atractides sp. selected all three substratum types (Figure 6a-c).
Most copepod species preferred sandy substrata, while immature copepodites and the cyclopoid Eucyclops serrulatus proximus also selected rocks and bryophytes (Figure 6df). Relative abundance of crenic invertebrate guilds, labeled as sedentary, ectoparasitic, and mobile on rocks (top), sand (middle), and bryophytes (down); the abundance values were transformed as follows: log(value + 1); numbers in orange squares: springs with a higher number of sedentary forms; numbers in green squares: springs with a higher number of mobile forms; sampling codes from 1 to 31, as in Table 1.
Sedentary forms recorded higher relative abundances compared to the mobile ones on rocky and sandy substrata (they reached higher abundances on rocks in 55% of springs and on sand in 65% of springs). Mobile groups had higher numbers in bryophytes in 77% of all sampling points.
The standardized index of preferences (B values > 0.33) depicted similar trends, with sedentary forms selecting rocky and sandy substrata in 22% and 64% of all springs, respectively. Mobile forms, however, showed preferences for sand and bryophytes in 64% and 51% of all springs, respectively. Ectoparasites were better represented in bryophyte substrata in terms of relative abundance and frequency ( Figure 5).

Habitat Preferences for Water Mites and Copepods
The habitat preferences for Hydrachnidia, which were calculated using the standardized selection index (B), showed that most water mite genera selected bryophytes, while Lebertia sp. and Atractides sp. selected all three substratum types (Figure 6a-c). These results must be treated with caution, since habitat selection was solely based on values of the standardized selection index that were higher than the threshold of 0.33, without a test of the null hypothesis of equal use for the three habitats. Moreover, the number of individuals is naturally small in these groups of organisms-in our case, it ranged from 250 to 350 in all sampling springs. These variations could also be due to the natural variability and the sampling effect. However, our findings present a trend of preference for bryophytes in the case of water mites and for sand in the case of copepods.
(c) (f) Figure 6. (a-c) Treemaps of habitat preferences for Hydrachnidia (left column) and (d-f) Copepoda (right column) according to the standardized selection index (B > 0.33 representing habitat selection; larger rectangles depict taxa with higher percentages of springs where rocks, sand, or bryophytes were preferred; taxa abbreviations are as in Table 2).  Table 2).
Most copepod species preferred sandy substrata, while immature copepodites and the cyclopoid Eucyclops serrulatus proximus also selected rocks and bryophytes (Figure 6d-f).
These results must be treated with caution, since habitat selection was solely based on values of the standardized selection index that were higher than the threshold of 0.33, without a test of the null hypothesis of equal use for the three habitats. Moreover, the number of individuals is naturally small in these groups of organisms-in our case, it ranged from 250 to 350 in all sampling springs. These variations could also be due to the natural variability and the sampling effect. However, our findings present a trend of preference for bryophytes in the case of water mites and for sand in the case of copepods.

Discussion
Our data showed the presence of the three analyzed substrata-rocks, sand, and bryophytes-in various percentages in all 31 sampling springs. Substratum heterogeneity was identified as an important driver of crenic invertebrate diversity in many studies, together with other abiotic factors, such as temperature or altitude [69][70][71]. Physico-chemical parameters were generally uniform in the 31 springs sampled in 2018, in accordance with the literature [23,27].
Amphipoda was the most frequent group in the study area, as in the majority of other studies dealing with crenic invertebrates [16,23,33,35,[72][73][74]. Other researchers also indicated Crustacea as the dominant group, but with Copepoda and Ostracoda [27,69]. Dipterans and, in particular, chironomids also recorded high relative abundances, similarly to other findings in the literature [13,70,75]. Other dominant groups mentioned in other studies were Trichoptera, Platyhelminthes [23,70], Gastropoda [74], Plecoptera, Coleoptera, and Ephemeroptera [13], which were also present in our data. Bonettini and Cantonati (1996) [76] found that Plecoptera, Ephemeroptera, and Trichoptera usually dominated in rheocrene springs, while Bivalvia, Nematoda, and Chironomidae did so in helocrene ones. However, no differences were observed in taxonomic compositions in the spring types that we considered, which was probably due to the dominance of rheocrene springs in our dataset.
Our data identified bryophytes as the preferred substrate for four groups, which were all insect taxa: Coleoptera, Diptera (Chironomidae and other families), Plecoptera, and Trichoptera. Rocks were selected by Ephemeroptera, while Copepoda preferred sand. Three groups recorded preferences for all three substratum types: Amphipoda, Platyhelminthes, and Ostracoda. Our results are difficult to compare with those of similar research due to the high heterogeneity of habitat classifications used in the literature. Bottazzi et al. (2011) [27], for example, reported different taxa and richness depending on the sampling techniques-benthic traps (Chironomidae, Ostracoda, other Diptera, crenophilous Harpacticoida, and Gastropoda), moss washing (crenophilic Harpacticoida, Ostracoda, Plecoptera, and Chironomidae), and drift tubes for groundwater fauna (crenophilous Harpacticoida, Chironomidae, and Plecoptera). Similarly to the trends depicted by our results, Bottazzi et al. (2011) [27] found that bryophytes supported an increased number of invertebrate groups. Other habitat categories were used by Dumnicka et al. (2007) [35], who divided spring substrata into fine versus coarse sediments. Amphipods (Gammaridae) and Gastropods (Bythinellinae) were abundant in both substratum types, while Platyhelminthes, Chironomidae, and Ostracoda prevailed in coarse substrata, and Trichoptera prevailed in fine sediments. Since no bryophytes were considered by Dumnicka et al. (2007) [35], the only common finding with the present study was the presence of amphipods in all habitat types. In their research, Reiss and Chifflard (2018) [33] considered spring invertebrates' preferences for 17 habitat types in deciduous versus coniferous forests. Their findings indicated coleopteran preferences towards moss habitats, as in the present study; however, they identified terrestrial forms. The highest number of taxa described by Reiss and Chifflard (2018) [33] preferred coarse particular organic material (CPOM), making comparisons with the present study difficult. Sun et al. (2020) [13] identified spring invertebrate indi-cator taxa in a review of 249 springs. The classes used by Sun et al. (2020) [13], such as groundwater-dependent species or stenothermal species, cannot be matched in our study.
Spring invertebrates' relative abundances were split into three guilds with respect to their ability to disperse from the crenic habitats: sedentary, ectoparasites, and mobile. No clear trends appeared in our data regarding the selection of habitat types by one particular guild. Higher abundances were recorded for sedentary forms in rocky and sandy habitats; these results were supported by the standardized selection index values. This could be due to the particular adaptations of the invertebrates that were characterized as sedentary in our study, i.e., copepods have spindle-like bodies that enable them to move in interstitial environments. Ectoparasites and mobile forms appeared to select bryophytes, which was probably due to the shelter that such an environment could provide. A small number of other studies divided crenic invertebrate communities into similar guilds. Williams and Williams (1998) [77] suggested that vagile insects (Plecoptera, Ephemeroptera, and Coleoptera) usually prevailed in springs that were recently subjected to glacial activity, while limnephilous organisms (Gastropoda, Bivalvia, Platyhelminthes, Amphipoda, Oligochaeta, Trichoptera, and Chionomidae) typically dominated in permanently flowing cold-water springs located in regions that were largely unaffected by recent glacial activity. Glazier (1991) [78] also separated insect and non-insect taxa, such as amphipods, based on water quality. Our data do not confirm such hypotheses, which is probably because the term "vagile insects" might refer to terrestrial forms as well, which were not considered in the present study. Similar research from the literature also refuted these trends [70,79].
The highest numbers of Hydrachnidia taxa and individuals were recorded in rheocrene springs (sampling point 5) and rheocrene cave springs (sites 6, 7, 13, 14, 17, and 18), which is comparable with similar findings from crenic habitats in Italy [68]. The elevated diversity of the taxa found in our samples (eight genera) and the relatively low number of water mite individuals compared to other groups could be explained by the feeding habits of the groups: parasitic larvae and predatory adults and nymphs [69,80]. The water mites' preference for bryophytes could be explained by an easier location of their prey, which are usually dipteran larvae of both chironomids and other families, in this environment [65]. Our results are in accordance with data from the literature, where higher abundances of Hydrachnidia were recorded in the mosses of small rheohelocrene springs (2000 individuals/square meter) compared to the gravel substrata of rheocrene habitats (where only 600 individuals/square meter were found) [68].
The copepod species identified in our data were found in similar studies on groundwater habitats [27,40,41,81]. Sun et al. (2020) [13] included most of them in the crenic bioindicator invertebrate list under the "groundwater-dependent species" category. Our results showed a higher preference of copepod species for sandy substrata. Similar data were found in the literature for cave or groundwater copepod species. Korbel et al. (2019) [82], for example, proved in a laboratory microcosm that Harpacticoida and Cyclopoida copepods preferred sand and gravel over clay, but showed no preference between gravel and sand. The harpacticoid ElaphoIidella sp. was found at high frequencies in the interstitial habitats of the alluvial plains of Slovenia [83].

Conclusions
The present study provides the first data on the habitat preferences of crenic invertebrates in Romania, in an aquatic environment that has been poorly investigated in our region. For this purpose, 31 springs were sampled in the Apuseni Mountains in the Romanian Carpathians with three types of substrata (rocks, sand, and bryophytes).
Four insect groups exhibited preferences for bryophytes (Coleoptera, Diptera, Plecoptera, and Trichoptera), while Amphipoda, the most abundant group in our samples, had equal preferences for the three habitat types.
The sedentary guild of invertebrates selected rocks and sand more frequently, while ectoparasites (Hydrachnidia) and mobile forms preferred bryophytes. Copepods displayed preferences for sandy habitats.