The 158 ha Nectandra Reserve (10°11'N, 84°31'W, elevation 1100 to 1200m) has an east-west axis along the 2 km arm of the L-shaped property. The vegetation is primary humid premontane forest (Holdridge Life Zone System 1967), with 98% forest canopy that occupies three-fourths of the property. The remaining one-fourth is a naturally regenerating forest (post 1980) on former coffee and Dracaena plantations. Two permanent creeks and four intermittently wet drainage streams cross the property to drain into the Balsa River. A network of narrow footpaths allows access to much of the mature forest interior and also its entire perimeter. A second smaller network of surfaced trails (gravel, paving stones and asphalt) surrounds the visitor facilities on the east end of the reserve.

Nectandra is subclassified as tall-statured montane cloud forest based on biologic and hydrometeorologic variables (Bruijnzeel et al. 2010). It has an average rainfall 3000-3500mm y^-1 and an estimated 80% fog saturated days. During the wet season (May–Oct), cloudy, overcast mornings with torrential afternoon rain are typical. During the wetter season (Nov- Feb) continuous rain with strong, wind-driven, heavy mist may last several weeks to months. Intermittent sunny to overcast days are mainly seen during the dry season (Mar-Apr). Volcanic basalt rocks, layered above a base of lava and volcanic agglomerates, cover much of the reserve. This thick matrix of volcanic ash form the clayey, vertical walls of many small canyons and eroded gullies.

The vascular plant diversity was determined from 30 permanent survey plots (10m x 20m) randomly distributed over the entire property (unpublished data). All the vascular plants in each plot were identified, tallied and tagged for monitoring. Of the total of 918 trees, at least 128 distinct species were identified. The most dominant plants in order of decreasing abundance included tree ferns Alsophila firma, Cyathea schiediana, Alsophila imrayana), non-ferns Guettarda poasana, Ocotea tonduzii, Conostegia oerstediana, Elaegia auriculata, each with species density averaging at least two individuals per plot. The 26 genera with the largest trees include Guarea (range 47–104 cm in diameter), Dussia, Hedyosmum (44-100 cm), Ficus (47-80 cm), Ocotea (42-76 cm) , Guatteria (60-75 cm), Paquira (40-73 cm), and Billia (53 – 76 cm) .

Between 2007 and 2009, D.N. made two separate surveys of mainly understory bryophytes at Nectandra. A third effort (2010) was made in conjunction with a tropical bryophyte course, taught by D.N and E.K. It was attended by eleven participants in the Bryophyte Study Group from seven countries, eight of whom contributed to the collection.

Haphazard floristic sampling of mosses, liverworts and lichens was carried out within 4m of 10 km of trails (equivalent sampled area of 4 ha) on all microhabitats ≤2m in height, including soil, soil banks, streams, rocks, tree trunks, branches, twigs, living leaves. The bryophyte growth on all surfaces is typically dense and entangled. One to five gram-size clumps or mats of candidate specimens were removed from the hard or pendant surfaces (bark/trunk, rock/asphalt) and placed in pre-numbered packets for evaluation and identification. The morphology of bryophyte species were examined in the laboratory with Zeiss dissecting and compound microscopes. With mixed clumps, the species of interest was teased away from the main clump and placed in an individual micro-packet for identification. Multiple micro-packets containing different, segregated, individual species of interest were prepared for each clump/leaf collected. Each outer packet hence contained the remainder of the clump and the associated micro-packets of the voucher specimens for the herbarium. The micro-packets in each main packet were differentiated by letters of the alphabet. The collection (see Supplementary material list A) was deposited at the Jepson Herbarium at the University of California at University of California, Berkeley. At the time of this report, 75% of the collection have been accessioned and the sample information are accessible online through www.ucjepson. Only bryophyte specimens with sufficient information and identification, completed to the level of genera were included in this report. To evaluate the completeness of the collection, we calculated the projected number of species, using the species accumulation curve (the function “specaccum”) and the species pool richness estimator (the function “specpool”) in the R-package “vegan”, from the accessions that were identified to the species level (Oksanen et al. 2016).

The identification of the specimens followed the key and description provided in Gradstein et al. 2001a and relevant monographs and literature therein. The nomenclatural information was verified with the TROPICOS database (www.tropicos.org).

There were a total of 1083 specimens examined at the genus level (Suppl. material 1). We identified 28 moss families (74 genera, 394 specimens), 24 liverworts (75 genera, 671 specimens) and 3 hornworts (3 genera, 18 specimens) families.

Only a subset of 488 specimens were identified with confidence to the species level (Suppl. material 2), in part due to the lack of complete monographic work in various taxonomic groups. Of these specimens, we found 188 species from 44 families (112 species of mosses, 74 species of liverworts, and 2 species of hornwort. The complete number of species was expected to be between 231-412 species (Fig. 2). However, since the majority of specimens were only identified to the generic level and difference in species identification between sites, we used the genus-level data for the subsequent analysis (Fig. 2)

Figure 2.  

Individual sample-based species accumulation curve from the bryophyte specimens from Nectandra, using the rarefraction method. The light blue band represents 95% confidence interval.

Fig. 3a shows the composition of the mosses. Eight families accounted for 58% (43/74 genera) of the diversity. These 8 families, listed in the order of decreasing family size, were Dicranaceae, Hookeriaceae, Pilotrichaceae, Hypnaceae, Neckeraceae, Bryaceae, Meteoriaceae, Sematophyllaceae, each composed of 4-8 genera. In contrast, a single liverwort family Lejeuneaceae with 40 genera accounted for 55% (40/75) of the total genera (Fig. 3b). The 3 families of hornworts were represented by a single genus each (data not shown).

a

b

Figure 3.

Composition of bryophyte genera by families at Nectandra Reserve, (A) moss families (74 genera) and (B) liverwort families (75 genera)

We classified the forest understory into 6 types of microhabitats: saxicolous, epiphyllous, corticolous, terricolous, aquatic rocks (submerged occasionally during heavy rain), and lignicolous. The first five categories were almost exclusively found in the forest understory. The last category, lignicolous, consisted mainly of exposed dry fence posts on the northern boundary adjacent to a treeless deforested ranch. It is the only habitat that receives full sun (Table 1).

Table 1 shows the analysis of 152 genera of bryophytes for their distribution in the above 6 habitats. Almost half of the mosses and liverworts were generalists (found on two or more substrate types) and half were specialists (found on a single type of microhabitat). More specifically, 34 of 74 (46%) genera are moss generalists and 35 of 75 (47%) are liverwort generalists. The remaining majority of taxa (40 genera each) of bryophytes are specialists, occupying unique niches in the understory (Fig. 4).

Figure 4.  

Distribution of bryophytes on six substrates

The corticolous bryophytes dominated with 112 genera collected off trunk, branches and twigs (Fig. 4). Together with the 47 epiphyllous and 28 lignicolous taxa, there were 187 genera of epiphytic bryophytes (109 of liverworts, 75 of mosses and 3 hornworts) growing on living or dead trees. In comparison to botanic substrates, only 85 taxa were collected off soil and rocks, with the mosses outnumbering the liverworts almost two to one (55 mosses to 30 liverworts). Hornworts were collected in equal number on soil and rock substrates. Only two genera of liverworts (Plagiochila and Cephaloziopsis) and one of moss (Sematophyllum) were found on boulders occasionally submerged during the heavy rainy season and flooding (Table 2).

We made detailed comparisons of the bryological diversity, at the level of genera at Nectandra with 4 other locations previously mentioned in the Introduction. Table 2 shows the locations, geophysical and climatic conditions of all 5 sites and tabulates the total number of bryophyte genera, as well as the numbe of genera in common or different from Nectandra. Compared to Nectandra, Monteverde showed lower number of bryophyte genera (47 mosses and 50 liverworts) but 92% were similar to those at Nectandra. Los Robles, on the other hand, has about the same number of mosses (72 genera) and few liverworts (50 genera), but many more were different (17% of both) from those at Nectandra. Santa Elena 95% of its mosses and 81% liverworts genera in common with Nectandra, with 17- 19% being different. The most intriguing site comparison is Cocos Island, where the 100% of 32 moss genera and 16 of liverworts coincide with those of Nectandra. However, the area of collection on Cocos Island was much smaller (0.08 ha) in comparison to the other study sites. Table 3, Table 4 detail shared and distinct bryophyte genera amongst the five sites.

From the accessions that we can determine with confidence, a total of 189 species were identified. This current number of species is far from complete, as suggested by increasing slope of the species accumulation curve and the discrepancy from the projected 213-412 species in Fig. 2. The remaining species are possibly among the undetermined specimens, or unexplored habitats, such as canopy or fully submerged habitat.

For practical reasons, we concentrated on the bryophytes accessible without climbing and included living as well as fallen trees. We did not estimate the proportion of specimens associated with freshly fallen trees. Our survey yielded a total of 152 genera of mosses, liverworts and hornworts at Nectandra, with a comparable number of hepatic and moss genera, although the number of genera per family differed markedly for the two groups (Fig. 3). The number of moss genera was evenly distributed across the families, whereas the hepatics showed a single dominant Lejeunaceae with 40 genera. Four liverwort families — Lejeuneaceae, Plagiochilaceae, Lepidoziaceae, Metzgeriaceae — accounted for 50% of the total 1083 specimens, whereas the largest 5 moss families (Sematophyllaceae, Dicranaceae, Calymperaceae, Hookeriaceae, and Polytrichaceae) represented only 17%. The dominance of these families are not surprising, because they are among the most species- and genera-rich families in the tropics (Goffinet and Shaw 2009). Lejeuneaceae, for example, consists of almost 100 genera, all of which are predominantly distributed in the tropics (Gradstein 2013). Two dominant pleurocarpous families (Sematophyllaceae and Hookeriaceae) are still subject to active taxonomic revisions (Hadenas 2012, Pollawatn 2008), and their number of genera could be still changing.

While there was an almost equal number of genera collected for mosses and liverworts, the ratios of specimens per genus was not — 394/74 for mosses vs. 671/75 for liverworts. This difference can be accounted quantitatively and qualitatively by Lejeuneaceae. Not only were there 40 genera in that single family, 68 % of the genera were generalists, growing on 2 of the 6 substrates tested, hence the large number of total liverwort specimens encountered. At the same time that there was a higher generic diversity in Lejeuneaceae, the number of individuals per genus in this family is lower — 7.5 compared to the average 10.9 for liverworts in the remaining genera and 10.7 for mosses in general. Given that we employed haphazard floristic sampling, the number of specimens should reflect more or less the abundance of bryophytes in the garden. Such a difference in the specimen numbers of mosses and liverworts highlights the tropical nature of bryophyte distribution, in which liverworts tend to be more common than they do in the temperate zone (Pócs 1982).

Our survey at Nectandra yielded 152 genera compared to 123 for Los Robles, 98 reported for Monteverde, 77 for Cocos and 36 for Santa Elena Peninsula. Given that the methodology and the surveyors differed at each site, it is not possible to make statistically meaningful comparisons. Nevertheless, it is useful to know the outcome of a coarse comparison among the five sites (Tables 2, 3, 4). The bryophyte richness at Los Robles is comparable to that at Nectandra, although they differ by 21 genera, the highest number of non-overlapping taxa among the 5 sites. Santa Elena Peninsula, not unexpectedly, was the least similar to Nectandra, due to the lower elevation and hotter and drier climate. Intriguingly, the bryodiversity on Cocos Island overlaps exactly with that of Nectandra at the genus level. Unfortunately, the survey on Cocos Island was done in a very small area (0.8 ha) and therefore difficult to compare with more certainty. Lastly, our nearest neighbor Monteverde has 90% of the taxa in common with Nectandra, with 9 distinct families. Our results are consistent with a similar but higher diversity at Nectandra than that of Monteverde. It is noteworthy that just in the Lejeuneaceae alone, 40 genera were collected at Nectandra, compared to 27 at Monteverde and 23 at Los Robles. Several families of mosses and liverworts were absent from Nectandra, but found in Monteverde and Los Robles. The absence can probably be attributed to our incomplete sampling from the canopy and other habitats or sampling missed in our survey. Sphagnum (Sphagnaceae), for example, only occurs in bogs with poor drainage and acidic water, which is not found in Nectandra. However the number of hornworts taxa at Nectandra was by far the highest among all the sites. Anthocerotaceae was present at Nectandra, Los Robles and on Cocos Island. Dendrocerotaceae was found at Monteverde and Nectandra whereas Notothyladaceae was unique to Nectandra.