Finding the forgotten gems: revisiting the butterflies of Matheran after 125 years with introduction to novel colour barcode for depicting seasons and activity of the Indian butterflies

Abstract We present here an updated checklist for the butterflies of Matheran, Maharashtra, India, an eco-sensitive zone, with identification remarks for locally rare or very rare butterflies. This is the first dedicated checklist for butterflies of Matheran after 125 years. A total of 140 species of butterflies were recorded belonging to six families. Amongst them, 15 species were either listed under Schedule I, II or IV of the Indian Wildlife (Protection) Act, 1972. We also list the habitats of the species along with the data for their activity at the time of recording the observation. We propose a uniform colour code system for representing season and activity for the Indian butterflies. Examples of colour barcodes are provided with the images of rare and very rare butterflies. The lack of abundance data is a limitation of the study for which we propose long term monitoring with dedicated efforts.


Introduction
Butterflies are an ideal taxonomic group for ecological studies of landscapes (Thomas and Malorie 1985) and their value as indicators of biotope quality is being increasingly recognised because of their sensitivity to minor changes in micro-habitat, particularly to the luminosity (Kremen 1992). Further, the butterflies are good biological indicators of habitat quality, as well as for the general health of the environment (Larsen 1988; Kocher and Williams 2000;Sawchik et al. 2005). Long-term diversity studies could, therefore, indicate the health of the habitat and ecosystems therein.
Here, we provide a checklist for butterflies of Matheran surveyed between the years 2011 and 2019. Ours is the first dedicated checklist for the butterflies of Matheran after Betham (1894). He listed 78 species of butterflies, combining the list of sixty butterflies provided by Smith (1882) and the list of butterflies recorded by him between April and May 1892. Padhye et al. (2013) provided a list of 27 butterflies from Matheran, while compiling the checklists for the butterflies of Northern Western Ghats, which was far from complete when compared to that given by Betham (1894). Further, the data on the habitat and seasonal turnover for butterflies of Matheran are particularly lacking from all these studies. Our checklist is accompanied with data on habitat, seasonal turnover and behavioural observations taken at the time of recording the species. We provide a novel coloured barcode approach for indicating the season/s and types of behaviour which could be used for all Indian butterflies. Representative colour barcodes are provided with the images of rare and scheduled species.

Study Area
Matheran (18.9866°N 73.2679°E, 772 m a.s.l., WGS 84) is a small hill station located in Karjat Tehsil of Raigad District in the Indian State of Maharashtra (Fig. 1). It is spread over an area of 7 sq. km. Matheran literally means forest on the top of the mountains. Geologically, it is a basaltic mesa separated from the main escarpment of Western Ghats by the low lying plains of Konkan and is an example of regressive erosion (Pascal 1988). Matheran gained the status of an Eco-Sensitive Zone (ESZ) in 2003 from the Ministry of Environment, Forest and Climate Change, Government of India [S. O. 133 (E)]. The ESZ of the Matheran comprises an area of 214.73 sq. km. All types of industrial, developmental and vehicular activities are restricted by this governmental order, making Matheran unique amongst hill stations of Asia. It experiences a cooler climate throughout the year (23.2°C mean annual temperature) compared to the surrounding low lying area and experiences heavy rainfall during the monsoon (4073 mm mean annual rainfall). The landscapes of Matheran are represented by open or forested laterite plateaus, hill-slopes, dense valley forests, non-perennial streams, manmade lakes, clearings near forest paths and human habitation. The flora of Matheran is represented by tree species found in mid elevation type wet evergreen forest (Ramesh et al. 1997), dominated by Memecylon umbellatum, Syzygium cumini and Actinodaphne lanceolata (Birdwood 1886, Ramesh et al. 1997. The plateau also hosts species like Carallia integerrima, Glochidion lanceolarium, Olea dioica, Garcinia indica and Carissa carandas (Birdwood 1886). The area also shows the presence of many endemic species of orchids, grasses and other herbaceous plants (Kothari and Moorthy 1993).

Field Survey and Data Collection
The area was visited in all the three seasons, namely summer (Feb-May), monsoon (Jun-Sept) and winter (Oct-Jan) throughout the year from September 2011 to March 2019. Intermittent observations were taken between 06.00 hrs and 17.00 hrs for around three days a month. The butterflies were observed in all possible habitats at six localities and on Study area with its location in Maharashtra, India. Sampling sites are shown in green filled circles. Additionally, the survey was conducted on two trails, Neral-Dasturi Road (pink line) and Neral-Matheran Railway (green line). two trails in and around Matheran (Table 1) (Vattakaven et al. 2016). To ascertain the identity of butterflies, photographs were taken and identifications were made with the keys provided by Evans (1932), Wynter-Blyth (1957, Kunte (2000), Kehimkar (2008), Kehimkar (2016) and Bhakare and Ogale (2018). The classification and nomenclature follows Kehimkar (2008), Van Gasse (2013 and Varshney and Smetacek (2015). The local status of the butterflies was decided, based on the number of records as very rare (≤ 5 records), rare (between 5 and 10), not common (between 10 and 20), common (between 20 and 50) and very common (> 50). This status does not correlate to the entire geographical distribution status of a corresponding species. The habitat, occurrence and behaviour of butterflies were noted and photo documented. The photo documentation was made with Nikon d500, d3200 and Cannon EOS 70d, Sony HX 100v digital cameras. The species were noted along with the date and location.

Data Analysis
Based on the occurrence data, a species accumulation curve (SAC) was prepared in R (R Core Team 2020) using the SpecAccum function in vegan (Oksanen et al. 2019). Expected (mean) species richness was calculated using the data collected from eight sites (Table 1). Further, the occurrence data of the species were analysed for calculating Similarity-Richness difference-Species replacement simplex (SDR Simplex) using SDRSimplex (a stand-alone computer programme) (Podani and Schmera 2011). Ternary plots were plotted using NonHier platform of SYNTAX 2000 (Podani 2001). The number or percentage of the species recorded per family, during each season, at each site was calculated in Microsoft Excel 2007 and visualised using pie and bar charts.

Preparation of Colour codes
The colour codes (Table 2) were prepared for easy and uniform representation of seasons and various behavioural activities of the Indian butterflies. Summer, monsoon and winter were given basic red, green and indigo colours in the CMYK scheme. These colours also correspond to temperature shifts in the seasons from hotter to cooler weather conditions. For combination of seasons, the corresponding combination of colours was used. Colours were mixed online through Color Mixer platform of Color Designer (https://colordesigner.io/ color-mixer). Grey colour represents the occurrence of the species in all seasons. All other colours were selected from the RGB scheme for it provides a wider range of colours. These colours were selected in such a way that they represent the correponding activity, for example, brown for mud puddling, honey colour (orange palette) for nectaring, amber colour for tree sap feeding etc., except basking which is represented by magenta.  Table 2.
Colour scheme for colour barcodes with CMYK and RGB ratios and HEX numbers.
Finding the forgotten gems: revisiting the butterflies of Matheran after ...

Species Richness
The SAC gained a plateau and standard deviation for species richness declined from 97.75 ± 17.07 to 141.0 ± 0.0 as the number of sights increased from one to eight, predicting sufficient efforts to record all the species found in the area (Asym = 146.42, xmid = 0.58, slope = 3.60) (Fig. 2). A total of 140 species belonging to six families have been observed and identified during the entire period of the study (Fig. 3, Table 3). The family Lycaenidae with 46 species (32.86%), followed by Nymphalidae with 43 species (31.43%), were amongst the most species-rich families in the area. Species belonging to the family Hesperiidae (25 species), Pieridae (14 species) and Papilionidae (10 species) were amongst other common species found in the area. The range of Cheritra freja (Common Imperial) which was earlier recorded from Amboli, Sindhudurga, Maharashtra (15.9647°N, 74.0036°E) (Saji and Ogale 2020) is extended further north around 345 km linear distance (calculated on https://www.nhc.noaa.gov/gccalc.shtml). The family Riodinidae was represented by only one species namely, Abisara bifasciata (Double Banded Judy).       Family-wise species composition pie of butterflies of Matheran.
Finding the forgotten gems: revisiting the butterflies of Matheran after ...

Seasonal turnover
The maximum numbers of species (N = 125) were recorded during winter, while minimum numbers of species (N = 80) were recorded during the monsoon (Fig. 4). Maximum numbers of species for all the families were recorded during winter, except the family Hesperiidae for which the maximum numbers of species (N = 23) were recorded during the monsoon (Fig. 5). The species of the family Lycaenidae dominated the local butterfly species richness during the months of summer and winter with 36.05% (N = 31) and 34.40% (N = 43) of total species of butterflies recorded during respective seasons (Fig. 6).
Members of the family Nymphalidae shared fairly equal percentages during all seasons. The percentage of the papilionids was the lowest during all seasons.  Family-wise percent species richness per season.

Spatial turnover
Members of the family Nymphalidae and Lycaenidae dominated the species diversity at all the sites studied in and around Matheran. Members of the family Lycaenidae were particularly present in higher numbers at Charlotte Lake while those of Hesperiidae were particularly present in higher numbers at Garbett Point (Fig. 7). The Similarity-Richness difference-Species replacement simplex for all the families indicated high similarity, although with different patterns tending towards perfect nestedness ( Fig. 8a-e, Suppl. material 1). Similarity was the highest for the family Nymphalidae (70.58%) with 78.22% of relativised strict nestedness (nestedness without considering the effect of species replacement) and lowest relativised beta diversity of 29.42%. Relativised strict nestedness was the highest (85.67%) for the family Hesperiidae with a similarity of 65.91% and beta diversity of 34.10%, while relativised nestedness (nestedness considering the effect of species replacement) was the highest (93.56%) for the family Pieridae. Similarity of species composition between the sites was the lowest (49.10%) for the family Lycaenidae with the highest relativised richness difference (31.99%) indicating more site specific species composition for the members of the family Lycaenidae, unlike the members of other families.

Activity of butterflies
No seasonal activity pattern could be observed (Table 3, Table 4). Most of the species were observed while mud puddling, basking or feeding on the nectar. Other common activities included feeding on bird droppings, tree sap, animal waste (other than that of birds) and/or animal carcasses. Site-wise percent species richness for each family

Locally rare and scheduled species
Our list contains 15 such species which are scheduled under the Wildlife (Protection) Act, 1972 of India (Table 5). Out of these, seven species were found rarely during the survey. Additionally, 20 species, which are not scheduled under the act, were observed rarely or very rarely during the survey (Figs 9, 10  Bibasis sena (Moore, 1865) (Fig. 9a). Habitat and activity: The species was observed in forested patches while nectaring.
Common name: Tamil spotted flat.
Identification remarks: Similar to common spotted flat, but UPF has semi-transparent white spots separated from each other. Markings on UPH indistinct or absent. Antennae chequered, club white in male, white at base only in female. Wingspan 45-50 mm.
Habitat and activity: The species was observed in forested patches while nectaring.

Identification remarks: Outwardly diffused broad white or bluish-white band on UNH.
Female has an additional spot on UPF. UN paler, inner half has greenish gloss. Wingspan 45-55 mm.
Habitat and activity: The species was observed in plains and undulating terrains while either mud puddling, basking, nectaring or feeding on carcass.
Identification remarks: A prominent row of black spots always towards the inner edge of the white band on both sides of HW. UPF white cell streak divided into four parts. Wingspan 60-70 mm.
Habitat and activity: The species was observed in forested patches while mud puddling or basking.
Habitat and activity: The species was observed in forested patches while mud puddling, basking or nectaring.
Common name: Brown king crow.
Identification remarks: Similar to Common Crow, but UN of either wing has no spots. All wings bordered with series of marginal and sub-marginal white spots. Male has a short, oval, dark band on UPF. UPH has greyish scales on apical half and pale-yellow scent scales patch. Wingspan 85-100 mm.
Season: Summer and winter.
Habitat and activity: The species was observed in forested patches while mud puddling, basking or nectaring.
Common name: Double branded crow.
Identification remarks: Similar to Common Crow, but male has two parallel brands on UPF; female has two similar faint streaks near inner edge on UPF. Wingspan 95-105 mm.
Habitat: The species was observed in forested patches while mud puddling, basking or nectaring.
Identification remarks: Pale greenish-yellow, wide central band on both sides. Large pale green spot near FW apex on both sides. Wingspan 60-75 mm.
Habitat: The species was observed in forested patches while mud puddling or basking, feeding on tree sap, animal waste or carcasses.
Habitat and activity: The species was found at forest edges while mud puddling or basking or feeding on tree sap, carcasses, animal waste, bird droppings or rotten fruits.
Habitat and activity: The species was observed in forested patches while mud puddling, basking or nectaring.
Finding the forgotten gems: revisiting the butterflies of Matheran after ...
Habitat and activity: The species was observed in forested patches while nectaring.
Appias libythea Fabricius, 1775 (Fig. 13f). the months of summer when they could be observed in dark, shady habitats. We were, however, unable to determine the cause of the high number of hesperiid observations during the monsoon and this needs a detailed behavioural study.

Spatial Turnover
The patterns for the diversity of butterflies of Matheran are very similar to those of the California Channel Island Birds and Vanuatu Birds, mentioned by Podani and Schmera (2011). High overall similarity for the entire butterfly diversity (Suppl. material 2) and familywise similarity between the sites (Fig. 8a-e) indicate the possibility of very stable diversity in the area with very low emigration to, or immigration from, surrounding areas. However, a detailed study from surrounding areas would be required to confirm this fact. The high overall similarity between the pairs of study sites (N = 28) also suggests a higher percentage of habitat generalist species surveyed in and around Matheran.

Colour coding
This novel approach is expected to improve the representation of the data for seasons and activities of the Indian butterflies. We encourage adding more activities and unique colour codes to make this system more universal, uniform and reader friendly. We also recommend its use while uploading records on open databases, such as Butterflies of India (Kunte et al. 2020) and iNaturalist (https://www.inaturalist.org/) for conveying information regarding the seasons and activities of butterflies.

Conclusions
A total of 140 species of butterflies belonging to six families were recorded from Matheran, India. This list includes 77 new records for Matheran. We observed a strong seasonal variation in butterfly diversity. The maximum diversity (N = 125) of butterflies was recorded during winter, while the least (N = 80) during monsoon. A high similarity of butterfly species composition was observed between the pairs of sites studied, tending towards perfect nestedness. This also emphasises the fact that the butterfly diversity in the region is quite stable and chances of emigration to, or immigration from, surrounding regions are very low. A strong seasonal gradient for activity patterns was not observed; however, we did observe a 'pocket effect' of dry season on butterflies. Butterflies during the dry season tend to aggregate near damp and shady places. Further, we introduce a novel barcode system for denoting seasons and activities of Indian butterflies and hope that this will help butterfly biologists to concisely and effectively present the data.
Heritage Library for making rare old manuscripts readily available online. We are grateful to Dr. Thomas Vattakven and India Biodiversity Portal for helping us upload raw data of the project and providing the URL for citation. NM thanks Manas Modak for helping prepare the raw dataset in Darwin Core Format through his excellent skills of programming in java. NM thanks Shruti Paripatyadar for introducing him to SDR simplex and its uses. We thank Rohan Bhagat for helping us prepare the map of the study site. MS and SS also thank Abhinav Nair, Gargi Geedh and Tejas Mehendale for helping them variously. We thank reviewers and subject editor for their invaluable comments which helped improve the manuscript. We are grateful to the editorial board and the journal for providing a generous waiver on article processing charges upon our request. Finally, we thank our families for keeping up the working environment at home amidst these chaotic COVID-19 situations.

Author contributions
MS and SS conducted the field survey. NM did data analysis. MS, SS and NM conceptualised and developed the colour code. MS, SS and NM wrote the manuscript.

Conflicts of interest
Authors declare no conflict of interest.