Naming and recognizing plant species
During all interviews, anglers identified by using official botanical or folk names a total of 16 plant species from the questionnaire (for the list of species and the number of all answers with relevant information, see Table 2). An angler correctly named 1–10 plant species of the total 24 (mean: 4.6), recognised, but not called by name an addition of 3–14 species (mean: 7.4), while was unable to recognise a total of 4–20 plants (mean: 11.8). A total of 45 names were used by anglers (Table 3). Scientific or folk names of 7 plants emerged regularly (more than 10 total mentions), which includes Phragmites australis (70 identifications by the name of the plant), Nymphaea alba (68), Lemna minor (62), Nuphar lutea (43), Iris pseudacorus (33), Trapa natans (29), and Salvinia natans (14). In the case of morphologically highly similar congeneric species pairs, anglers did not differentiate Sparganium emersum and Sparganium erectum, Myriophyllum spicatum and Myriophyllum verticillatum, while the informants clearly distinguished Ceratophyllum submersum and Ceratophyllum demersum. Most anglers treated together, and did not distinguish the following floating or rooted aquatic plant species (commonly named ‘hínár’, ca. pondweed in Hungarian): Potamogeton nodosus, Stuckenia pectinata, Trapa natans, Myriophyllum spicatum, Myriophyllum verticillatum, and Salvinia natans, while Ceratophyllum submersum and Ceratophyllum demersum were also called ‘hínár’, but as mentioned above, was clearly distinguished by anglers.
During the interviews, we heard a total of 66 answers of 20 botanically dialectical names, but after understanding the mental representation of anglers by using the control questions of the questionnaire, and also considering the available Hungarian botanical, ethnographical, and ethnobiological literature, 62 (88.5%) of these answers related to 16 botanically dialectical name were considered as likely accurate, especially when the dialectical name was used consistently (see Table 3 for details). The names ‘lily’ (liliom) and ‘water lily’ (vízililiom) were used equally for Nuphar lutea, Iris pseudacorus and Nymphaea alba; these are widely used names for both I. pseudacorus and N. alba in the country [44, 45]; therefore, the answers were accepted as dialectical names for the other two species. Five reported folk names, as ‘kecskeköröm’: (ca. goat-nail) for Trapa natans, ‘balinhínár’ (ca. asp’s pondweed) for Myriophyllum spicatum/verticillatum, ‘rucarence’ (ca. duck’s bladder-wort), ‘panyola’ and ‘kákonya’ (untranslatable names) for Salvinia natans need further investigations due to the limited number of reports, the origin, distribution, and other variants of these names. (See all plant names and likely accurate answers also in Table 3).
Despite the fact that they could not name them, the interviewed anglers could recognise and had knowledge of many of the plant species, mostly which are salient, or somehow related to their fishing activities (Table 2, Figs. 2,3). More than half of the interviewed anglers could answer to at least two of questions 2–4 (see Data collection in Methods) in the case of Ceratophyllum demersum (62 correct answers without the name of the plant), Butomus umbellatus (43), Potamogeton nodosus (41), Glyceria maxima (38), and Myriophyllum spicatum (38). At least half of the interviewed anglers did not know the name and were unable to recognise Marsilea quadrifolia (71), Juncus bufonius (69), Hippuris vulgaris (65), Ceratophyllum submersum (65), Sparganium emersum (60), Sparganium demersum (60), Lemna trisulca (58), Stuckenia pectinata (42), Hydrocharis morsus-ranae (42), Nymphoides peltata (40), and Trapa natans (38). None of the anglers could name Lemna trisulca, Potamogeton nodosus, Stuckenia pectinata, Glyceria maxima, Hippuris vulgaris, Juncus effusus, Juncus bufonius, and Marsilea quadrifolia.
FEK of protected status and consumption by humans and wild animals
A total of 38 anglers (52%) reported that Nymphaea alba is legally protected. Only six of the anglers (8%) mentioned the protected status of Trapa natans, whilst no other legally protected plant species in the questionnaire (Nymphoides peltata, Hippuris vulgaris, Salvinia natans, Marsilea quadrifolia) were mentioned during the interviews. To our question about the possible human consumption, six anglers (8%) reported that Trapa natans is edible, while no other plants were mentioned for possible human consumption. For the question ‘Do you know about any animals (including fish) which consume this plant?’, 48 anglers (66%) reported that Ctenopharyngodon idella regularly consumes fresh or mature sprouts of Phragmites australis, while other relevant plant consuming fish, bird, or mammal species were reported scarcely (bird species – 19 times; other fish species – 18 times; mammal species – 8 times, mainly reporting that animals occurring around freshwater habitats consuming ‘pondweeds’ in general), and except for reporting the Ctenopharyngodon idella, half of the anglers (n = 36) did not mention any animals which consume any of the plant species specifically.
Anglers miscellaneous perceptions on freshwater plant species and their surroundings
While anglers formed neutral opinions of most of the floating or rooted plant species, some of the species repeatedly suffered from rather negative judgements: Trapa natans - ‘long branches are needed in the front of the fishing piers against them (see Fig. 3)’; ‘if I can reach from my pier, I always pull out the rosettes from the water’; ‘it is an aggressive plant, it can even germinate from 2 metres under water’; ‘this plant is not legally protected for sure, or if yes, whoever protects it is not normal’, Ceratophyllum demersum – ‘it always gets stuck on the hook, I hate it’; ’it destroys the boat’s engine, and squeezing the clutch’, Salvinia natans – ‘if it multiplies, you can collect it all day long’, aquatic plants in general – ‘the fish taste of mud mainly because of this plant, it needs to be exterminated’; ‘it needs to be exterminated, it becomes entangled in the fishing line’.
During the interviews, regardless of the questions asked, some of the anglers reported information about the current status of vegetation (e.g. ‘There are too dense patches of Phragmites australis here, it has to be cut at the winter’), some of them observed and verbalised long-time changes in the vegetation and density of selected plant species (e.g. ‘Trapa natans was less widespread 30 years ago then now’) or in habitat management (e.g. ‘The canal is very turbid, so dredging is urgently needed here’).
While only one of the interviewees knew the name of Nymphoides peltata, 31 other anglers (43% of the total interviewees) had seen the plant before, and 23 anglers (31% of the total interviewees) reported about specific sites of the species in the country, mostly Lake Tisza, which represents the greatest population of the plant in the country. Seven interviewees even at Lake Velence remembered this species from Lake Tisza. Three anglers reported negative experiences about struggling by boat in a dense population of N. peltata (e.g. ‘There is so much of this plant in the Lake Tisza that you can barely move in the plant mass with the boat’).
Factors affecting FEK
Our categorization of plants, including characteristics of appearance, relatedness and commonness, was confirmed by the results of discriminant analysis. The average proportion of correctly identified categories was 0.83 (range: 0.8-1, Fig. 4) and only the two species in the fourth category (non-salient, non-fishing related, rare) were assigned to the third category (salient, non-fishing related, rare). The proportion of named or recognized species is significantly lower in categories with decreasing relatedness to fishing and commonness (R2 = 0.74, ### (intercept), β = 0.96, t = 11.23, P < 0.001; ##X, β=-0.41, t=-4.2, P < 0.001; #XX, β=-0.75, t=-6.53, P < 0.001; XXX, β=-0.91, t=-6.11, P < 0.001). However, by estimating the relationships between the same proportion and each of the characteristics (R2 = 0.74), we found significant associations in relatedness and commonness (β = 0.34, t = 3.75, P = 0.001 and β = 0.41, t = 4.21, P < 0.001, respectively), but not in salience (β = 0.16, t = 1.09, P = 0.29). According to the results of our linear models, some of the components of fishing experience, namely the years spent with fishing (β = 0.003, t = 2.55, P = 0.013) and the number of visited locations (β = 0.01, t = 3.07, P = 0.003) showed significant differences in the proportion of named or recognized plant species. Furthermore, only Lake Velencei had significant association with species number (β=-0.1, t=-2.23, P = 0.029), where anglers recognized less plant species compared to Keleti Main Canal. These results were also confirmed by the model selection (Table 4,5).
Table 4
Comparison of truly competitive models of the proportion of plant recognition and its predictors in the best subset (ΔAICc < 4) of models, ordered by AICc values.
Model No. | Predictors | Df | logLik | AICc | ΔAICc | Weight |
1 | number of fishing locations + years of fishing | 4 | 52.51 | -96.43 | 0.14 | 0.29 |
2 | age + number of fishing locations + years of fishing | 5 | 52.97 | -95.03 | 1.55 | 0.14 |
3 | number of fishing locations + site | 5 | 52.44 | -93.98 | 2.60 | 0.09 |
4 | number of fishing locations + job type + site + years of fishing | 8 | 55.90 | -93.52 | 3.05 | 0.07 |
5 | number of fishing locations + residence + site | 6 | 53.30 | -93.30 | 3.27 | 0.06 |
6 | age + number of fishing locations + residence + site + years of fishing | 8 | 55.76 | -93.23 | 3.34 | 0.06 |
7 | frequency of fishing + number of fishing locations + years of fishing | 7 | 54.41 | -93.07 | 3.50 | 0.05 |
8 | years at the current location + number of fishing locations + site | 6 | 53.14 | -92.98 | 3.59 | 0.05 |
9 | age + fishing competition + number of fishing locations + years of fishing | 6 | 53.06 | -92.83 | 3.74 | 0.05 |
10 | age + years at the current location + number of fishing locations + years of fishing | 6 | 53.03 | -92.76 | 3.81 | 0.05 |
11 | age + fishing for consumption + number of fishing locations + years of fishing | 6 | 53.02 | -92.75 | 3.82 | 0.05 |
12 | age + number of fishing locations + residence + years of fishing | 6 | 52.99 | -92.68 | 3.89 | 0.04 |
Table 5
Model-averaged parameter estimates (β), standard errors (SE), and variable importance (I) for the predictors of plant recognition. Levels of categorical variables shown in parentheses were averaged separately.
Predictor | β | SE | I |
number of fishing locations | 0.014 | 0.004 | 1 |
years of fishing | 0.003 | 0.001 | 0.8 |
age | -0.001 | 0.001 | 0.39 |
site (Keleti) | | | 0.32 |
site (Látókép) | -0.045 | 0.039 | |
site (Velencei) | -0.099 | 0.041 | |
job type (blue-collar) | | | 0.07 |
job type (retired) | -0.016 | 0.039 | |
job type (white-collar) | 0.036 | 0.033 | |
residence | 0.025 | 0.037 | 0.16 |
fishing frequency (daily) | | | 0.05 |
fishing frequency (monthly) | 0.049 | 0.067 | |
fishing frequency (weekly) | 0.087 | 0.062 | |
fishing frequency (yearly) | 0.023 | 0.087 | |
years at the current location | 0.001 | 0.001 | 0.1 |
fishing competition | 0.015 | 0.035 | 0.05 |
fishing for consumption | 0.011 | 0.036 | 0.05 |
The fact that anglers better recognized aquatic plant species alongside Keleti Main Canal compared to Lake Velencei was also confirmed by Dunn’s test (Z = 2.7, P = 0.02). Any other comparisons for the categorical predictors were statistically non-significant, however, we were interested in the differences in other fishing related experiences of anglers among fishing sites, among job types and in the frequency of fishing. We found that retired anglers are fishing significantly longer than people in blue-collar (Z = 3.2, P = 0.02) or white-collar jobs (Z = 3.2, P = 0.04) and similarly, anglers at Lake Látókép or at Keleti Main Canal have started fishing earlier in their life compared to Lake Velencei (Z = 2.81, P = 0.007; Z = 3.09, P = 0.006, respectively). There were no significant differences in the rest of the comparisons.
The 24 plant species represent different levels of their recognition, therefore, they contributed unequally to the knowledge of anglers. We applied principal component analysis on the raw values (named, recognized and not known) given for each species based on the answers of the 72 interviewees. The average of Cronbach’s alpha was 0.729 and the θ was 0.77 when we included all species in the analysis. However, excluding some of the variables resulted in higher reliability of the analysis. After removing 11 items out of 24, Cronbach’s alpha became 0.777 and θ was 0.79. We extracted six principal components (PC1 to PC6), explaining 83.43% of the total variance, for further analysis. PC1 represented the recognition of T. natans and S. natans, PC2 was highly contributed by Iris pseudacorus, Nuphar lutea and T. natans, whilst PC3 was characterized by Sparganium emersum, S. erectum and I. pseudacorus (see more details in Table 6). Using the six principal components for k-mean clustering, we identified 3 groups (Fig. 5). Anglers in group 1 are fishing significantly shorter than in groups 2 or 3 (Dunn’s test, group 1 ~ group 2: Z=-2.27, P = 0.03; group 1 ~ group 3: Z=-3.67, P < 0.001), are visiting the current fishing location more recently than people in group 3 (Z=-3.13, P = 0.005). The proportion of named or recognized plant species was decreasing among the groups (group 1 ~ group 2: Z=-2.45, P = 0.01; group 1 ~ group 3: Z=-5.42, P < 0.001; group 2 ~ group 3: Z=-3.39, P = 0.001). Groups were not statistically different in other comparisons, as well as, we did not find significant differences in age and the number of fishing locations regularly visited among the groups.
Table 6
Variable loadings on principal components. Negative numbers represent lower values in the raw data, thus, they indicate that anglers are more likely to name or recognize the species.
| PC1 | PC2 | PC3 | PC4 | PC5 | PC6 |
Nymphoides peltata | -0.111 | 0.1356 | 0.169 | -0.0579 | -0.6289 | -0.1032 |
Nuphar lutea | -0.3058 | 0.4991 | -0.2301 | 0.6623 | -0.0283 | -0.0354 |
Lemna minor | -0.0818 | 0.161 | 0.0095 | 0.2666 | 0.0416 | -0.1168 |
Potamogeton nodosus | -0.2145 | -0.0345 | 0.063 | -0.2067 | -0.3276 | 0.2175 |
Hydrocharis morsus-ranae | -0.2396 | -0.2134 | -0.2546 | -0.1065 | -0.0734 | 0.5461 |
Sparganium emersum | -0.0794 | 0.0776 | -0.5234 | -0.2998 | 0.0249 | -0.3356 |
Sparganium erectum | -0.0794 | 0.0776 | -0.5234 | -0.2998 | 0.0249 | -0.3356 |
Trapa natans | -0.603 | -0.4113 | 0.2817 | 0.0578 | -0.0973 | -0.5019 |
Ceratophyllum demersum | -0.0848 | 0.1334 | -0.0439 | 0.008 | 0.1079 | 0.1239 |
Juncus bufonius | -0.0287 | 0.0257 | -0.0341 | -0.0709 | 0.0184 | -0.0197 |
Iris pseudacorus | -0.2817 | 0.551 | 0.4292 | -0.4538 | 0.4063 | -0.0016 |
Salvinia natans | -0.4943 | -0.2702 | -0.1462 | 0.0863 | 0.3995 | 0.2605 |
Butomus umbellatus | -0.2759 | 0.2891 | -0.1174 | -0.1767 | -0.3766 | 0.261 |