The material was hand-collected by Thomas Wesener and Jan Philip Oeyen in the Halimun-Salak National Park during the dry season, between the 9th of September and the 7th of October 2015 (29 days) with the help of local guides and an Indonesian research assistant.

Three different habitats and sites were explored:

1. Mt. Salak and Sukamantri in the eastern area around the volcano,
2. Cidahu below the mountain, and
3. Cikaniki in the centre of the National Park, where the Research Station was located.

See Fig. 1.

Figure 1.  

Map of the Halimun-Salak National Park and location of the three collecting sites Mt. Salak and Sukamantri, Cidahu and Cikaniki, modified after Lund and Rachman (2018).

These areas are exposed to different degrees of human influence, which is reflected in their vegetation.

Habitat 1, Mt. Salak, had only little natural vegetation left, which was dominated by banana plantations and Casuarina forests.

Habitat 2, Cidahu, was semi-natural. It was a former deforested area and has now been regenerating for several years. On the one hand there were some freshly regrown indigenous rainforest parts, on the other hand some alien species and even some neophytes like bananas (Musa) and Australian Casuarina had started to spread within the floral compositions.

Habitat 3, Cikaniki, was the only near-natural area. It had primary rainforests and only selective logging was found at some places.

Every single collection event in the three habitats received its own identification number (labelled “HAL XXX” for Halimun-Salak National Park).

All specimens were collected by hand and preserved in 96% ethanol. Quantitative sampling was not possible as sieves and Winkler extractors were retained in customs.

Every habitat received a different amount of collecting effort of 10 hours per day per person. Night catches added 4 more hours per worker to the workload. Habitat 3, Cikaniki received the highest amount of sampling hours and was the only collection site with night catch events.

The specimens were sorted and all the individual sample glasses got voucher labels (e.g. HAL45-02). Each label stated the voucher number, the HAL number, the exact location with the GPS data and the date when the sample was collected.

For a better and quicker identification, photos of millipedes were taken with a Leica Z6APO Q Imaging System (Q Imaging, Canada; www.qimaging.com) together with the programme Auto-Montage Pro 5.03.0061 from the company Synoptics (Cambridge, Great Britain 2006; http://www.synoptics.co.uk). The pictures were edited in Adobe Photoshop CS2 9.0.2.

Sometimes, large amount of juveniles were collected at a single collection event (probably nests of freshly hatched eggs). To avoid a strong bias caused by these random findings, only three specimens of those nests were included in the data analyses.

It was not possible to identify the juveniles to species-level (or even genus- or family-level). Thus, unidentified juveniles were not included in the data analyses at the species- or generic-level. The collected material will be shared between Museum Koenig, Bonn, Germany and LIPI, Bogor, Indonesia.

As no determination keys for Indonesian millipedes exist, numerous encountered species might be undescribed, and most of the described Indonesian Myriapoda genera and species are in need of revision. Therefore, we chose the morphospecies approach. In several cases, especially for numerous Polydesmida and Geophilomorpha, a trustworthy determination to genus level was not possible. In a few cases, a determination to family level was not possible for the Spirostreptida. Those specimens were labelled as “Spirostreptida Family 1 Genus A spec. 1”. For a complete list of determinations, refer to the supplementary material (Suppl. material 1).

For diverse groups outside our area of expertise, the identifications were verified by international experts. The scolopendromorphs were, for instance, identified by Arkady Schileyko (Moscow Lomonosov State University, Moscow), while the polydesmidans were identified by Sergei Golovatch (Russian Academy of Sciences, Moscow) and Weixin Liu (South China Agricultural University, Guangzhou).

The morphospecies determination for the genus Otostigmus (Scolopendromorpha, Chilopoda) was problematic. Only identification to species-groups (after Lewis 2010) was possible, likely leading to an underestimation of the actual number of Otostigmus species in our samples.

After the identification of all specimens to class, order, family, (morpho-)genus and morphospecies, they were sorted in a chart in Microsoft Excel 2010 (Microsoft Corporation) for the data analyses. Symphyla were removed from all except the class-level statistics, as morphospecies determination was not attempted.

For the specimen-level analyses, juveniles (except when more than 3 were collected in a swarm, see above) were retained, whereas they were removed from the analyses at the morphospecies-level. 769 specimens (226 Chilopoda and 543 Diplopoda) were available for the specimen analyses, while 617 specimens could be incorporated in the morphospecies analyses.

A species accumulation analysis was performed with the software “environment for statistical computing and graphics R” version 3.3.0 (03.05.2016) developed by the R-Core Team in 1993 (https://www.r-project.org) (Ihaka and Gentleman 1996). Four species accumulation curves were obtained.

Of a total of 980 collected specimens, excluding a high number of juveniles, 796 Myriapoda samples from the Halimun-Salak National Park were analysed and 617 samples were identified to morphospecies (Figs 2, 3, 4). Nearly two-thirds (543 samples) were determined as Diplopoda. A little more than one-quarter (226 samples) of the remaining Myriapoda were identified as Chilopoda. A small fraction of 4% (27 samples) belonged to the class Symphyla, which was excluded from the study.

a

b

Figure 2.

Living specimens of Diplopoda and Chilopoda, photographed during the dry season of 2016 at Cikaniki, Halimun National Park, Java, Indonesia. All photographs by Jan Philip Oeyen.

a: Spirostreptidae, Harpagophoridae, Thyrophygus sp.  
b: Scolopendromorpha, Scolopendra subspinipes Leach, 1815  

a

b

c

d

Figure 3.

Living specimens of Diplopoda (Glomeridesmida, Glomerida, Sphaerotheriidae, Chordeumatidae) photographed during the dry season of 2016 at Cikaniki, Halimun National Park, Java, Indonesia. All photographs by Jan Philip Oeyen.

a: Glomeridesmida, Glomeridesmus sp., female  
b: Glomerida, Hyleoglomeris sp.  
c: Sphaerotheriida, Castanotherium sp.  
d: Chordeumatida, family and genus indet.  

a

b

c

d

Figure 4.

Living specimens of Diplopoda (Polydesmida) photographed during the dry season of 2016 at Cikaniki, Halimun National Park, Java, Indonesia. All photographs by Jan Philip Oeyen.

a: Cryptodesmidae  
b: Platyrhacidae, two different species  
c: Paradoxosomatidae  
d: Paradoxosomatidae  

The samples revealed seven different families of the orders Scutigeromorpha, Lithobiomorpha, Scolopendromorpha and Geophilomorpha.

The majority of the Chilopoda collected at Halimun-Salak (145 individuals; 64%) belong to the order Scolopendromorpha. The second most represented order are the Geophilomorpha with 31% (70 individuals). The Lithobiomorpha are represented by nine specimens (4%) and the Scutigeromorpha are represented by only two specimens (1%).

About 24 morphospecies (or alternatively species-groups in the case of Otostigmus) of Chilopoda were collected in the Halimun-Salak National Park (Fig. 5), with one species each of Scutigeromorpha and Lithobiomorpha. The centipede species diversity was, on the other hand, the highest in the Geophilomorpha (67%; 16 species), followed by the Scolopendromorpha (25%; 6 species). The most common centipede and myriapod species in general was Otostigmus aff. rugulosus, with 76 specimens representing 33% of all Chilopoda specimens and 12.3% of all the Myriapoda samples.

Figure 5.  

Percentage distribution of the Chilopoda species into the four orders.

A total of 18 different families and nine orders (Polyxenida, Glomeridesmida, Glomerida, Sphaerotheriida, Siphonophorida, Chordeumatida, Polydesmida, Spirobolida and Spirostreptida) were present in the sample. The most dominant millipede group was the Polydesmida (Table 1). Nearly two-thirds of all Diplopoda (69%; 374 specimens) belonged to the Polydesmida. The superorder Juliformia was represented by the orders Spirobolida (8%; 44 individuals) and Spirostreptida (7%; 37 individuals). Of the Nematophora, only the order Chordeumatida (1%; 4 specimens) was found in the Halimun-Salak National Park, whereas more than 85% of all collected Diplopoda belonged to the Eugnatha. Of the four orders of the Colobognatha, only one order, Siphonophorida (6%; 31 specimens), was found. The infraclass Pentazonia was represented by the orders Glomeridesmida (4%; 25 samples), Glomerida (1%; 6 samples) and Sphaerotheriida (3%; 18 samples). With only four individuals (1%), the Polyxenida was the least represented group in the Halimun-Salak National Park.

The dominance of the Polydesmida is also reflected in the morphospecies analysis, albeit to a lesser degree (Fig. 6), with 21 (45%) of all observed species. With 10 species (22%), the Spirostreptida showed the second highest species diversity. The Sphaerotheriida also showed high species diversity with seven morphospecies (15%). Glomerida, Siphonophorida and Spirobolida were represented by two species each (4%). The orders Polyxenida, Glomeridesmida and Chordeumatida were present with one morphospecies (2%) each in the Halimun-Salak National Park. The high diversity of the Polydesmida is further reflected at the family- and genus level. Six families, viz. Cryptodesmidae, Haplodesmidae, Trichopolydesmidae, Polydesmidae, Platyrhacidae and Paradoxosomatidae, were present, with the vast majority of genera (8 of 14) and morphospecies (13 of 21) belonging to the latter.

Figure 6.  

Percentage distribution of the Diplopoda at species-level.

The first curve (Fig. 7) represents all collected samples and species in the entire Halimun-Salak National Park, to determine if the results are meaningful enough to be compared to other Myriapoda faunal compositions. The three other curves (Fig. 8) represent the species accumulation for each habitat. The graph of the forest remnants and plantation (habitat 1) shows no significant indication of saturation (Fig. 8). The one for the former deforested and now regenerating area (habitat 2) seems to become slowly steady although it does not reach the saturation line. Only the amount of collected specimens from the primary rainforest (habitat 3) was enough to determine a saturation line in the species accumulation analysis. The species accumulation curve for all collected samples (Fig. 7) in the Halimun-Salak National Park shows a typical saturation curve with a quick increase of found species per specimen number, with the slope of the curve decreasing as samples were collected until reaching a certain saturation point.

Figure 7.  

Species accumulation curve for the Halimun-Salak National Park.

Figure 8.  

Species accumulation curve for the three different searching locations.