New records of Lophoproctus coecus Pocock, 1894 (Diplopoda, Polyxenida, Lophoproctidae) extend the range of the genus Lophoproctus

Abstract The geographic distribution of the genus Lophoproctus Pocock, 1894 has greatly expanded with new records of the species Lophoproctus coecus Pocock, 1894, together with the reassignment of a number of millipedes formerly identified as Lophoproctus lucidus (Chalande, 1888). Lophoproctus coecus was found to be the sole representative of the family Lophoproctidae in collections examined from Crimea and the Caucasian region. The species was also identified from Iran and Kyrgyzstan. Lophoproctus specimens collected in Italy by Verhoeff were reassigned as Lophoproctus coecus with the exception of one specimen of Lophoproctus jeanneli (Brölemann, 1910) from Capri. These data were combined with all available information from the literature to look at the pattern of distribution of the four species in the genus. The range of the genus Lophoproctus extends from Portugal to Central Asia. Lophoproctus coecus is widespread from Italy eastward, while the morphologically very similar species Lophoproctus lucidus is confined to France and northern Africa. The two species have a narrow overlap in the Alpes Maritimes region of France. Lophoproctus jeanneli has a scattered coastal distribution around the Mediterranean Sea. The troglobitic species Lophoproctus pagesi (Condé, 1982) has only been recorded from a cave on Majorca, Spain.


Introduction
Genera and species in the family Lophoproctidae Silvestri, 1897 have very similar morphology, with species adapted to an endogenous mode of life, being found in soil, caves and under stones. All Lophoproctidae lack ocelli, their integument lacks pigmentation and the 8 th antennal article is elongate. Species in the family also share the same arrangement of caudal trichomes and similar organisation of tergal trichomes (Fig. 1).
Characters used to determine genus and species of the family can be difficult to observe and include number and arrangement of antennal sensilla, number and arrangement of linguiform processes along the anterior margin of the labrum, structure of the telotarsus, leg setae, and tarsal spine. There are currently 5 genera: Lophoproctus Pocock, 1894, Lophoturus Brolemann, 1931, Ancistroxenus Schubart, 1947, Lophoproctinus Silvestri, 1948, and Alloproctoides Marquet & Condé, 1950. Two further genera Barroxenus Chamberlin, 1940 and Trichoproctus Silvestri, 1899, known only from single collections, are of uncertain status as they are inadequately described. Pocock (1894) established the genus Lophoproctus for a species collected from soil at Nervi in Liguria, Italy. As the species lacked ocelli he called it coecus ('blind' in latin). Previously in 1888, Chalande had described the species Pollyxenus (sic) lucidus from Palalda, Eastern Pyrenees, France, which he initially described as having ocelli. In 1894 Silvestri identified Polyxenus lucidus from Italy, then later the same year recognising that the specimens had no ocelli he moved the species into the genus Lophoproctus. He further suggested that L. coecus and L. lucidus were synonymous (Silvestri 1894b) as did Verhoeff some years later . Both Silvestri and Verhoeff collected widely throughout Italy (Silvestri 1894a, 1894b, 1952 and identified all lophoproctids they found as L. lucidus with the exception of those from Capri that Verhoeff mistakenly described as a new species Lophoproctus litoralis (Verhoeff, 1952). L. litoralis was later determined to be Lophoproctus jeanneli (Brölemann, 1910) (Condé 1969). Condé (1978) re-examined material from Verhoeff's collection from Isernia and Teramo, Zannone (Pontine Islands) and Sardinia, Italy, and noted that they differed from L. lucidus in that they had a different arrangement of sensilla on antennal article VI which was also more elongate. On the basis of Condé's description, Nguyen Duy-Jacquemin (1993) confirmed that these specimens were L. coecus.
In 1993, Nguyen Duy-Jacquemin redescribed L. coecus from syntypes from Nervi, Italy that together with specimens collected in Zannone, Italy by Verhoeff and from Rome, Italy by Silvestri, confirmed that L. lucidus and L. coecus were not synonymous.
Her paper clearly illustrates that the two species differ in arrangement of sensilla on antennal article VI (Fig. 2), structure of the median lobe on the anterior edge of the labrum, the number of ridges on the leg setae and the ratio of the length of the tarsal spine to length of the claw (Fig. 4). Initially described as a subspecies of L. lucidus, L. jeanneli is also found in the Mediterranean region of Europe but is easily distinguishable from other species of Lophoproctus by the presence of a denticle on the claw of the telotarsus.
The most recently described species Lophoproctus pagesi Condé, 1982, is a troglobitic species collected in caves on the island of Majorca, Spain. Specimens similar to L. pagesi have also been collected from caves in Portugal. L. pagesi differs from other species of Lophoproctus in details of the labrum, antennal sensilla and telotarsus. As well it has elongate antennae and legs typical of troglobitic species (Condé 1982, Nguyen Duy-Jacquemin 1993.   In this study I re-assessed specimens from The Zoological State Collection, Munich collected by Verhoeff in Italy and identified as L. lucidus. I also identified Lophoproctidae from Crimea, Caucasus, Iran and Kyrgyzstan in the collection of the Zoological Museum of Moscow. These data were then combined with details obtained from the published literature on Lophoproctus species to determine distribution of species in the genus.

Methods
The material examined for this study is lodged in the Zoological Museum of Moscow and the Zoological State Collection in Munich, Germany.
Specimens from the Zoological Museum of Moscow were examined and identified. All specimens were preserved in ethanol. These specimens were examined by light microscopy and scanning electron microscopy. For light microscopy, specimens were mounted on slides in Hoyer's medium, dried at 60 °C and examined with an Olympus CX 41 compound microscope. Scanning electron micrographs were obtained of selected whole specimens that were dehydrated in a graded series of ethanol, 80%, 90% and 100%, then air-dried. Specimens were then mounted on stubs using adhesive tabs, sputter-coated with gold and examined with a Philips XL20 scanning electron microscope. Photographs of whole specimens were taken with a Leica Integrated Stereomicroscope System comprising a Leica 205C microscope with a DFC425 camera and 5000HDI dome illuminator. Images were stacked using Leica Application Suite and enhanced using Adobe Photoshop CS6.
Specimens from the Zoological State Collection, Munich, are slide mounts in Canada Balsam made by KW. Verhoeff. The slides lack both date of collection and site habitat details. The slides were examined by light microscopy using an Olympus CX41 compound microscope. Due to the thickness of the slide mounts, they could not be examined at magnifications higher than 400×.
As no coordinates were available for most of the material examined, Google Earth was used to provide an estimate of geographical position for mapping purposes (a table of localities with coordinates is available in supplementary material). A map of the distribution of all species in the genus Lophoproctus was generated using SimpleMappr (Shorthouse 2010). New records determined in this study were included together with all known published records. Many records in the literature, especially those by Verhoeff, Silvestri and Tabacaru are questionable and these have been treated separately.

Results
Sixty collections of Polyxenida in the Zoological Museum of Moscow were examined and L. coecus identified in 15. In most cases less than 5 specimens were collected at a site. No other species of Lophoproctidae were found.
Twenty slides from the Zoological State Collection, Munich (ZSM/Myr. 20031594-612, 615) all labelled as L. lucidus were examined and 19 found to be L. coecus. Slide ZSM/Myr. 20031615 contained a whole mount of L. jeanneli.

Discussion
Lophoproctus coecus has previously been considered to occupy a scattered range within the Central Mediterranean region, but the results of this study indicate that the species is widespread throughout Europe particularly in Eastern Europe with its distribution extending into Central Asia. L. lucidus in comparison seems limited to the Southern France, as well as Morocco, Algeria and Tunesia in Northern Africa. The identification of L. jeanneli from Capri, Italy reinforces the Mediterranean coastal distribution previously noted by Kime and Enghoff (2011). It is of interest that species within the genus may overlap in their geographic distribution with both L. jeanneli and L. coecus being found in Capri and in the Alpes Maritimes region of France. Lophoproctus pagesi and L. jeanneli both occur on the island of Majorca, with L. pagesi restricted to caves while L. jeanneli was found in humid, sunny hilly areas. A number of identifications were unable to be checked. In the case of the Lophoproctus identified by Lignau as L. lucidus, specimens identified as L. coecus in this study were found at all 3 of Lignau's collection areas, indicating that it is most probable that the specimens collected by Lignau were in fact L. coecus. Prior to publication of Nguyen Duy-Jacquemin (1993), the difference between L. coecus and L. lucidus was not understood with Silvestri (1894b) considering the two species synonymous. Hence, until it can be confirmed, the identification of specimens of Lophoproctus as L. lucidus, from Romania  and Sicily (Silvestri 1903, Strasser 1970 must remain questionable. Lophoproctus has also been collected from Bulgaria but has not yet been identified to species (Stoev and Lapeva-Gjonova 2006).
The distribution map (Fig. 5) indicates very clearly that there is a big gap in our knowledge of Lophoproctus in Greece, the Balkans and Turkey. It is predicted that L. coecus does occur in these three locations, and that L. jeanneli may also occur in coastal regions. All species except the troglobitic L. pagesi have the ability to live in many of the habitats in these areas of Europe and Asia. Unfortunately there is limited information available to guide collection of these tiny millipedes. In most cases millipedes in this study were collected by hand collecting in the field, or by sieving of litter and/or soil followed by direct collection from a tray of sieved material (Fig. 6). Recently in Dagomys, Russia, L. coecus was collected from forest litter by funnel extraction (M. Potapov pers comm. 2014). Habitats from which Lophoproctus has been collected vary from maquis and forest litter, top layer of soil, cave floors, to under stones and logs, and in ants nests (Stoev and Lapeva-Gjonova 2006). In the case of L. coecus from Kyrgyzstan, the single specimen was collected from under stones on a dry grassy limestone slope, a similar habitat to that of the type collected by Pocock (1894), 'open hillside beneath stones'. As well as the above mentioned habitats, L. lucidus has been found in North African cities in city parks, in gardens and under pot plants and stones (N Akkari pers comm. 2014, , and near Algiers on the beach in accretions of both dry and damp marine plant material . Lophoproctus jeanneli has been collected in abundance from cracks in bricks and under dry stones on the ground in heavy shade in the Barcelona City Park  as well as on the sea shore (Verhoeff 1952).
Methods of dispersal have not been studied, but it is probable that the presence of L. lucidus in North African cities is due to anthropogenic activities dating back to the French colonial period as millipedes confirmed to be L. lucidus appear limited to Southern France and centres of French colonial activity in North Africa. Polyxenida are thought to passively disperse via the wind and incidental attachment to the feathers of birds. It is likely in the case of soil and litter dwelling lophoproctids that dispersal via wind is less common as they do not appear to be living in elevated situations in trees and bushes in contrast to Polyxenida from the families Synxenidae and Polyxenidae. However, limited methods of dispersal do not seem to have restricted the geographic ranges of species of the genus Lophoproctus, especially L. coecus.