Field identification of large-culmed bamboos

With care, identification of large-culmed woody bamboos within our survey regions was relatively simple, at least to genus. Of the large-culmed clumping bamboos (Fig. 1), the most frequently encountered was Bambusa vulgaris, with both the green and the yellow forms commonly planted in or near rivers, towns and villages, and we recorded it in the Makira, CAZ, Nosivolo and COFAV survey regions. The larger Dendrocalamus giganteus was less frequently observed. Of the non-clumping species, Valiha diffusa was easily recognizable within its range by its characteristic drooping culm tips, long lateral branches, stiff culm sheaths readily shed, and a thin whitish ring above the internodes following the shedding of the culm sheath (Fig. 2). Cathariostachys spp. often lacked major lateral branches, had persistent culm sheaths that were not readily shed, and often showed a pale ring under the internodes (Figs. 3–6). Arundinaria spp. had slimmer culms, with delicate culm sheaths often appearing torn or shredded, and a distinctive pattern of leaf growth (Figs. 7–10). The young emerging culm shoots of each genus were also highly distinctive (Figs. 1–10).

Figure 1.

The pantropical bamboo Bambusa vulgaris (left: green form; centre: yellow form) and the introduced Dendrocalamus giganteus (right). Photographs by Tony King and Maherisoa Ratolojanahary.

Local names for bamboos were very variable. Valiha diffusa was usually referred to as Vologasy or Volojatsy, although both names were used for other species in some places. Cathariostachys bamboos were usually called either Volohosy, Volo lagnana or Volobe, but in lowland areas Volobe was more often used for the green form of Bambusa vulgaris. The yellow form of B. vulgaris was usually called Volovanga or Volovazaha. Arundinaria bamboos were almost always called Volotsanganana, and we only rarely heard this name used for other bamboos. The use of local names varied from region to region, but also between local people in the same region—sometimes simply due to misidentincation of bamboos to the appropriate local name by less-experienced local guides. Names used for various smaller-culmed bamboos included Volohando, Volohoto, Voloandotra, Volomadinika, Volotami-hana, Volosodina, Volokitrana, Volovahy and Tsimbolovolo.

Distribution and elevation ranges of Valiha diffusa, Cathariostachys and Arundinaria

We made 1,343 georeferenced records of endemic large-culmed bamboos (Valiha diffusa, Cathariostachys and Arundinaria) across the seven survey regions. The majority of our observations (74%, n = 994) were made in and around the Ankeniheny-Zahamena forest corridor (CAZ), where we have long-term monitoring of Prolemur simus populations. One region had only one observation (Anjozorobe), the remaining five had between 43 and 129 observations each. Cathariostachys bamboos were recorded from Makira in the north to Midongy in the south, Valiha diffusa from Makira to the southern parts of the COFAV, and Arundinaria bamboos from Zahamena to the COFAV (Table 2; Figs. 11–14).

Valiha diffusa was only recorded at low elevations (9-576 m), and Arundinaria only at high elevations (1018–1667 m), while Cathariostachys bamboos were distributed across a wide range of elevations, from 53 to 1471 m (Table 2, Fig. 11). Low-elevation records of Cathariostachys were rare (Fig. 11), with only five of 710 records (0.7%) below 300 m, 12 (1.7%) below 400 m, and 65 (9.2%) below 600 m. Very high elevation records of Cathariostachys were also rare, with only three records (0.4%) over 1,300 m. Most records of Cathariostachys could therefore be considered mid-elevation, with 90.4% (n = 642) between 600 and 1,260 m.

Figure 2.

The Malagasy endemic bamboo Valiha diffusa in lowland deforested landscapes of eastern Madagascar. Photographs by Lucien Randrianarimanana and Hery Randriahaingo.

Table 2.

Geographic coordinates of the 47 survey sites, within seven survey regions, with elevation ranges at each site for endemic large-culmed bamboos Valiha diffusa, Cathariostachys spp. and Arundinaria spp.

continued

Figure 3.

Cathariostachys sp. near Makira in north-eastern Madagascar. Photographs by Lucien Randrianarimanana.

At the most northerly region surveyed, Makira (c. 14.8°S), there appeared to be a continuum of elevation records for Cathariostachys, from 340 to 1040 m (Fig. 11). At our more central survey regions, from 17 to 22°S (Zahamena, CAZ, Nosivolo, COFAV), there appeared to be two groups of elevation records, below 600 and above 800 m, with very few records between 600 and 800 m (Fig. 11). At the most southerly region, Midongy (23 to 24°S), most records were between 400 and 1000 m, with two below 200 m and none between 200 and 400 m (Fig. 11).

Figure 4.

Cathariostachys madagascariensis in the western Ankeniheny-Zahamena Corridor in eastern Madagascar. Photographs by Lucien Randrianarimanana and Tony King.

Culm diameters

Valiha diffusa had the most variable culm diameters at our sites, ranging from 1.7 to 9.5 cm (Table 3). Cathariostachys had the largest mean culm diameter of the endemic large-culmed bamboos (6.1 cm), and Arundinaria the smallest (4.0 cm). Culm diameters of the green form of the pantropical Bambusa vulgaris (measured at the Ambalafary site to the east of the CAZ) were larger than those of the endemic bamboos, ranging from 3.2 to 13.4 cm, with a mean of 9.58 cm (SD 2.04, n = 248).

There were some inter-site differences in culm diameters of the endemic bamboos (Table 3). For example, the Cathariostachys species at the lowland sites of Maherivaratra (Makira region) and Vohibe (Nosivolo region) had mean culm diameters one to 2.5 cm smaller than at the other, higher elevation, sites (Table 3). This is a statistically significant difference, when comparing each with the site with the next lowest mean Cathariostachys culm dbh measured by the same researcher (Maherivaratra with Ranomainty: z = 14.08, p<0.001; Vohibe with Zahamena: z = 8.80, p<0.001).

Table 3.

Summary statistics for culm dbh measurements for endemic large-culmed bamboos at some of the survey sites.

Figure 5.

Cathariostachys sp. in and around the Midongy du Sud National Park in south-eastern Madagascar. Photographs by Laingo Rakotonirina.

The mean culm diameter of Valiha diffusa was 0.6 cm lower in open areas than in areas of closed or semi-closed canopy cover at both lowland sites where this was measured, a statistically significant difference at both sites (Table 4).

Table 4.

Comparison of mean culm diameters (dbh) of Valiha diffusa under differing levels of canopy cover at two lowland sites in the CAZ survey region, using the z test for unmatched samples.

Figure 6.

Cathariostachys sp. (probably C. capitata) in the lowland forest fragment of Vohibe in eastern Madagascar. Photographs by Anjara Bonaventure, Anselmo Andrianandrasana and Tiana Ratolojanahary.

Distribution of large-culmed bamboos in Madagascar

By comparing the distribution of our field observations with georeferenced published herbarium specimen locations (Table 5), our study has resulted in a southern extension of the published range of Valiha diffusa (Fig. 12), and southern and northern range extensions for species of Cathariostachys (Fig. 13). We have also shown that Arundinaria occurs in remaining high elevation forests in the Zahamena National Park, the Ankeniheny-Zahamena Corridor, the Anjozorobe Corridor, and around the high-reaches of the Nosivolo River (Fig. 14). We recommend the collection of herbarium specimens from these sites to help resolve the taxonomic uncertainties surrounding the Malagasy bamboos currently included in Arundinaria (Dransfield 2000, 2003; or alternatively in Yushania and Thamnocalamus by The Plant List 2010).

Herbarium specimens are also needed to help determine species' distributions within Cathariostachys, as we were unable to differentiate the two recognized Cathariostachys species in the field, and also found no consistent pattern that might separate them based on elevation. It may be that there is only one, highly variable, Cathariostachys species, as suggested by Dransfield (1998), although the thinner mean culm diameter of Cathariostachys at our lowland sites of Maheri-varatra and Vohibe compared to our higher elevation sites is consistent with the descriptions of the two species given by Dransfield (1998). The species identity of Valiha in the Ankarana National Park in the far north of Madagascar also requires further study (Dransfield 2003). Although close to the type locality of Valiha perrieri, the large-culmed bamboos of Ankarana appear very similar to Valiha diffusa (L. Rako-tonirina and T. King, unpubl. data), and probably are indeed the latter (S. Dransfield, in litt. 2012).

Bamboo distribution is often affected by disturbance (Griscom and Ashton 2006; Gagnon et al. 2007; Olson et al. 2013). However, disturbance may affect different bamboo species differently. Cathariostachys madagascariensis appears to be found at higher densities in disturbed forests than in non-disturbed forests, perhaps indicating an evolutionary adaptation to regular cyclonic disturbance in the eastern rainforests (Olson et al. 2013), but does not appear to tolerate excessive disturbance, and does not persist in deforested areas far from the forest edge (Fig. 13; see also Olson et al. 2013). Valiha diffusa, conversely, appears to thrive in many deforested landscapes of lowland eastern Madagascar (Figs. 2 and 12; Dransfield 1998), and is probably more abundant now than when lowland forest cover was more extensive. However, although apparently more tolerant to major disturbance than Cathariostachys, V diffusa probably also has a limit to its tolerance of disturbance, with culm diameters reduced in heavily-disturbed areas where it is cut or burnt regularly (Dransfield 1998); a phenomenon common to several bamboo species (Franklin et al. 2010). We also found that V. diffusa culm diameters were lower in more open habitats, possibly due to higher rates of culm production in open areas (Gagnon et al. 2007), or perhaps due to environmental stress related to extreme micro-habitat variables such as moisture or light levels (Reid et al. 1991).

Table 5.

Herbarium specimen locations for Valiha spp., Cathariostachys spp., and Arundinaria spp., given by Dransfield (1998, 2003), georeferenced for this study.

Large-culmed bamboos and Prolemur simus

At least 29 species have been recorded as food plants of Prolemur simus, of which nine are woody bamboos, ten are other monocotyledons, and ten are dicotyledons (Table 6). The Poaceae family (including bamboos) makes up 45% of recorded food plants. At sites where quantitative studies have been undertaken, over 90% of the feeding time of P. simus is devoted to one or two of the large-culmed bamboo species Cathariostachys madagascariensis, Valiha diffusa, Bambusa vulgaris and Arundinaria sp. (Tan 1999; Mihaminekena et al. 2012; Randrianarimanana et al. 2012). Bambusa vulgaris may or may not be native to Madagascar (Dransfield 2003), and does not occur in natural forest in the country (this study). The native large-culmed bamboos of Cathariostachys, Valiha and Arundinaria are, therefore, likely to be the primary natural food plants of P. simus. Of these, Arundinaria species are restricted to high elevation forest, and appear to be less frequently consumed by P. simus than the more widespread Cathariostachys species and Valiha.

If Cathariostachys and Valiha can therefore be considered the principle food plants of P. simus, their distribution should be one of the major factors influencing the range of P. simus itself. The distribution map we provide for Cathariostachys (Fig. 13) does appear to correspond well to the potential current distribution of P. simus (Fig. 15) based on recent direct sightings and indirect feeding signs (as given by Rakotonirina et al. 2011) and unconfirmed local knowledge (as given by Dolch et al. 2010), with perhaps the notable exception of the Masoala peninsula in the north-east of the country from where there is currently no indication of the presence of P. simus.

Figure 7.

Arundinaria sp. in high elevation forest at Anjozorobe in eastern Madagascar. Photographs by Tony King.

Sub-fossil records of Prolemur simus are more widely distributed (Godfrey et al. 2004), occurring in areas where Cathariostachys species do not occur, at least in recent times (Fig. 15). However, Valiha does occur in at least one of these sub-fossil sites (Ankarana), and therefore may have been the primary food plant of the species in that area, and conceivably could still be so (L. Rakotonirina and T. King, unpubl. data). There is also an unidentified bamboo resembling Valiha in western dry forest near Morondava (Dransfield 2003), which may have been the food plant of P. simus when it existed in that part of the country. Further research into the historical distribution of large-culmed bamboos in Madagascar would be very helpful in establishing how eventual changes in bamboo distributions over time may have affected changes in P. simus distribution.

Table 6.

Recorded foodplants of Prolemur simus at five sites in eastern Madagascar, based on 1: Petter et al. 1977; 2: Meier & Rumpler 1987; 3: Tan 1999; 4: Tan 2007; 5: McGuire et al. 2009; 6: Ravaloharimanitra et al. 2011; 7: Rakotonirina et al. 2011; 8: Bonaventure et al. 2012; 9: Mihaminekena et al. 2012; 10: Lantovololona et al. 2012; 11: Randrianarimanana et al. 2012; 12: L. Randrianarimanana, unpubl. report 2011; 13: Andrianandrasana et al. in press.

Figure 8.

Arundinaria sp. in high elevation forest of the western Ankeniheny-Zahamena Corridor in eastern Madagascar. Photographs by Lucien Randrianarimanana.

Figure 9.

Arundinaria sp. in high elevation forest at Tsinjoarivo in eastern Madagascar. Photographs by Laingo Rakotonirina.

Figure 10.

Arundinaria cf. ambositrensis in high elevation forest at Ranomafana National Park in south-eastern Madagascar. Photographs by Tony King.

Figure 11.

Scatter graph of GPS points taken for Valiha diffusa (blue), Cathariostachys spp. (red) and Arundinaria spp. (green), by elevation and latitude.

Prolemur simus is known to have a very seasonal diet (Tan 1999), feeding primarily on young large-culmed bamboo shoots when they are available (Figs. 16–18), and switching to other bamboo parts when they are not. In Ranomafana National Park, P. simus feeds primarily on the young ground shoots of Cathariostachys madagascariensis during the early wet season months of November to February, then supplements its diet with bamboo leaves and branch shoots as the number of ground-emerging shoots declines, whilst during the drier months of June to October it feeds principally on the mature culm pith, still of C. madagascariensis (Tan 1999). At our study sites, the seasonality of the emergence of ground shoots of Cathariostachys madagascariensis was generally similar to that described by Tan (1999), as was that of Valiha diffusa, with the latter perhaps a month later than C madagascariensis. Other large-culmed bamboo species, however, did not follow the same seasonality, especially those in the genus Arundinaria and the yellow variety of Bambusa vulgaris. The presence of these bamboos may therefore influence the seasonality of P. simus feeding behaviour if ground shoots emerge when those of other species are not available. This is particularly apparent at the lowland Ambalafary site, where the P. simus group spends a lot of time feeding on the ground shoots of yellow Bambusa vulgaris during the dry, austral winter months from June onwards (T. H. Mihaminek-ena, unpubl. data; Fig. 19). Similarly, at the high-elevation Ranomainty site, ground shoots of Arundinaria appear to be sometimes eaten during the austral winter dry season, especially in drier years when Cathariostachys ground shoots are rarer (H.L.L. Randrianarimanana, unpubl. data).

Figure 12.

Distribution of herbarium specimens (black stars) and our field observations (gray diamonds) of Valiha diffusa, with approximate forest cover given in gray.

Figure 13.

Distribution of herbarium specimens (black stars) and our field observations (gray diamonds) of Cathariostachys spp., with approximate forest cover given in gray.