Dung beetle vicariant speciation in the mountains of Oaxaca, Mexico, with a description of a new species of Phanaeus (Coleoptera, Geotrupidae, Scarabaeidae)

Abstract An analysis of vicariant speciation of Geotrupes and Phanaeus (Coleoptera, Geotrupidae, Scarabaeidae) from the mountains of Oaxaca, Mexico is undertaken. The new species of dung beetle (Coleoptera: Scarabaeidae) from Oaxaca, Mexico, Phanaeus dionysius sp. n. is described. Photos of the habitus and a distribution map are provided. Phanaeus malyi Arnaud is revalidated. An updated key for the Phanaeus endymion species group and new localities are also presented. An updated key for the Geotrupes of Oaxaca and new locality records are also submitted.


Introduction
Recent collecting and taxonomic studies of dung beetles of the genera Geotrupes Latreille (Coleoptera: Geotrupidae) and Phanaeus MacLeay (Coleoptera: Scarabaeidae) in the mountains of Oaxaca have evidenced the existence of a repeated vicariant speciation pattern. This pattern seems to be located at present between the northern (Sierra Norte) and southern (Sierra Sur) mountain ranges of this state. An analysis of this vicariant speciation mechanism is presented, based on the great taxonomic similarity (sister species) shown by the Geotrupes viridiobscurus Jekel-Geotrupes pecki Howden and Phanaeus dionysius sp. n.-Phanaeus zapotecus Edmonds species pairs where the first species inhabits the northern and the second one the southern mountain systems.
In this paper Phanaeus (Notiophanaeus) dionysius sp. n. is also described, a species that inhabits dry oak scrub forest in the Mexican state of Oaxaca, between 1900 and 2200 m above sea level and belongs to the Phanaeus endymion species group. The number of Phanaeus in Mexico has been increasing during the last years. Since Edmonds and Zídek's monographic analysis in 2012, three new species have been described just in 2017 from Mexico ) and one more in 2018 from Ecuador (Arnaud 2018).
This new species brings the number of known Mexican Phanaeus species to 32 (see Edmonds and Zídek 2012, . Mexico seems to be a hotbed for Phanaeus speciation because more than half of the species (32 out of 58 described species; 34 if we also include the USA) are distributed in this country. As a comparison, Colombia registers eight species (Medina et al. 2001), Costa Rica also eight species (Solís and Kohlmann 2012), and Panama only four (Solís and Kohlmann 2012). This suggests that Phanaeus has invaded North America, probably since the Miocene (Edmonds 1994), and that Phanaeus found empty niches (where there were no other Phanaeini genera competing) and/or reduced local competition that promoted speciation. We consider therefore that the most probable local North American ecological equivalent, the Oniticellini, were outcompeted by them or perhaps they were poorly represented from the beginning. Philips (2016) considers the invasion of North America by the Oniticellini from Asia via Beringia to be an old event, most likely taking place during the Eocene, or perhaps even earlier during the late Oligocene or early Miocene. On the mainland, the only species of Attavicinus Philips and Bell, and two species of Liatongus Reitter, are known at present, living in restricted niches like ant debris, carrion, and fungi (Philips 2016). Philips (2016) considers necrophagy and mycetophagy to have evolved recently, therefore suggesting that the Oniticellini could have changed their feeding habits recently under a competitive pressure process exerted by the Phanaeini.
An updated key as well as new localities and commentaries for species of the Phanaeus endymion species group are also presented. We revalidate Phanaeus malyi Arnaud, a member of the aforementioned species group. New distribution records for Phanaeus endymion species group (Ph. bravoensis, Ph. halffterorum and Ph. huichol) are included. New distribution records for Geotrupes viridiobscurus and Geotrupes pecki also are added (G. pecki was last reported 40 years ago; Howden 1974), and an analysis of their distribution is presented. An updated key for Oaxacan Geotrupes also is presented.

Materials and methods
The superb personal collection of Julián Blackaller, in Soria, Guanajuato, Mexico, was reviewed; as well as the collection of the National Museum in Costa Rica and the Canadian Museum of Nature in Ottawa, Canada. Collections made by Arriaga-Jiménez as part of her studies on mountain dung-beetles, which are deposited in the entomology collection at the Institute of Ecology, Xalapa, Mexico, were also studied. Body measurements were made to the nearest 0.1 mm using an ocular micrometer with a Stemi DV4 stereoscope. Genital dissections and preparations were done following the techniques described by Zunino (1978). Genital structures were stored in microvials with glycerin.
The photos were taken by Alfonso Aceves from the Instituto de Ecología (IN-ECOL), Xalapa, Mexico, using a Canon T2i camera, extension tubes, a 100 mm macro, and an external Canon flash. The photos of Ph. malyi and Ph. pyrois were taken by Ángel Solís from the Museo Nacional de Costa Rica using an Olympus OM-D E-M5 digital camera. Susana Guzmán-Gómez from the Instituto de Biología, UNAM, Mexico, took the photo of the pygidium of Ph. zapotecus, using a Zeiss AXIO Zoom V16 microscope, a Zeiss AxioCam MRc5 camera, and the ZEN (Zeiss Efficient Navigation) multifocal technology programme. The photos of Ph. zapotecus and the Suppl. material 1 ( Figure S1) were taken by François Génier, from the Canadian Museum of Nature, Ottawa, using a Leica Z16 system and LAS software for image stacking.
The holotype, allotype, and two paratypes of Ph. dionysius are deposited in the Colección Entomológica (Entomology Collection), Instituto de Ecología, Xalapa, Mexico (IEXA). Further paratypes are deposited in: two paratypes (male and female) in the Canadian Museum of Nature, Ottawa (CMN); two paratypes (male and female) in the Julián Blackaller (JB) personal collection, Soria, Mexico; two paratypes (male and female) in the Patrick and Florent Arnaud (CPFA) personal collection, Saintry sur Seine, France; and two paratypes (male and female) in the Seção de Entomologia da Coleção Zoológica da Universidade Federal de Mato Grosso, Cuiabá (CEMT), Brasil. Diagnosis. Distinctly granulate male pronotal disk; sagittal furrow present on the female pronotum; unmodified sutural margin of the elytra; pygidium longer than wide. Its basal border forming a small indentation medially, usually all-black color.
Description. Holotype. Major male ( Fig. 1a-b). Length: 16.5 mm. Humeral width: 10. 9 mm. Body appearing dull shiny black with a faint blue luster to the unaided eye. Magnification reveals faint greenish cast along the ocular, pronotal, and elytral borders, on the abdominal surface and underside of femora rugose. Clypeus with two conspicuous median teeth; surface and frons bearing long, slender horn strongly curved over the pronotum. Pronotum with large, flat triangular disk (Fig.  1a), with a well-developed, small callosity on each side near anterior margin and with postero-lateral angles projecting caudally; lateral portions faintly asperate, with distinct punctures present only behind lateral fossae (×20); flat triangular surface disk densely, evenly, and coarsely granulate; granules extending onto posterolateral angles and becoming eroded near and along disk borders. Pronotum with obsolete basal fossae; anterolateral angles subquadrate, distinctly upturned and surface behind angles concave; pronotal midline present, faintly developed, more evident on anterior half; weak punctures along posterior pronotal margin; pronotal surface shagreen. Elytral striae fine, with small but well-defined punctures separated at regular intervals; intervals broad and faintly convex, evenly and faintly shiny, covered with minute punctures (×20); surface shagreen. Pygidium black with shagreen surface and obsolete punctures, glabrous; pygidium wider than long (Fig. 2a); basal pygidial margin forming a small triangular tooth medially (Fig. 2b); pygidial margin with a green cast. Protibia quadridentate. Lamella copulatrix as in Figure 3a; aedeagus similar to the Ph. endymion species group (Fig. 3b, d).
Female. Allotype (Fig. 4). Length: 16.3 mm. Humeral width: 10 mm. Body faintly shining black. Head with low, narrow trituberculate carina. Pronotum with a faint green lustre, evenly and densely covered with punctures, punctures becoming fainter on middle of disk; surface shagreen; with raised anteromedian trituberculate tumosity near anterior margin, tubercles equal in size and set in a more-or-less straight, transverse line; disk with distinct mid-longitudinal furrow, extending forward from posterior margin to about middle of disk, furrow more strongly sculptured than adjacent surface of disk. Pygidium with faint to distinct fine, sparse punctures.
Variations. Length: 12.6-18.7 mm. Humeral width: 7.9-11.1 mm. Pronotal disk of males may vary from black without reflections to having a green or red lustre. Minor male (Fig. 5): Similar to major male, except the cephalic horn is smaller and the posterolateral angles of the pronotum are reduced.   Etymology. Due to the fact that this species has been collected in association with Ph. damocles Harold, this new species (a noun in the nominative singular) is named after Dionysius II of Syracuse, one of the main characters alluded to in the moral anecdote of the "Sword of Damocles". Distribution and ecology. So far, this species is only known from San Pablo Etla in the Sierra Norte (Sierra de Ixtlán) in Oaxaca (Fig. 6), along the internal dry slope facing the Oaxaca Valley. It has been collected from altitudes of 1950 m to 2250 m. The dry deciduous oak forest where Phanaeus dionysius sp. n. was found is characterised by trees between five and ten meters tall. Abundant oak species are Quercus laeta Liebm. and Q. laurina Humb. and Bonpl., predominant species of this ecosystem, ranging from 1800 m to 2400 m altitude (Fig. 7). Other species dominating this forest in the sampling site are Q. glaucoides Mart. and Gal., Q. liebmannii Oersted., Q. rugosa Née, and Q. castanea Née, also found at higher or lower altitudes (J. Williams, CIIDIR-Oaxaca, pers. comm., Valencia-Ávalos and Nixon 2004). This dry deciduous oak forest shows a strong seasonality, when most trees lose their leaves for around four to five months between December and May. This new species is found next to the Oaxaca Metropolitan Area in a voluntary protected area, the San Pablo Etla Community Reserve "La Mesita".
Phanaeus dionysius sp. n. has been collected simultaneously in dung-baited traps with Canthidium quercetorum Kohlmann, Arriaga-Jiménez and Rös, Canthon humectus   to be a strictly mycetophagous species, Ph. dionysius has only been collected in dung, despite the presence of fungi-baited traps put in the forest. Interestingly, no big fleshy fungi (toadstools) were observed in this type of forest, only small "clavitos" (Lyophyllum), which could probably explain why this species does not exploit fungi.
Taxonomic relationships. Phanaeus dionysius sp. n. belongs to the Ph. endymion species group and due to its close taxonomic similarity discussed below is postulated to be the sister species of Ph. zapotecus Edmonds, 2006. Phanaeus dionysius will key out to Ph. zapotecus in the key of  and can be separated from it because it has long and slender pronotal posterolateral angles (Fig. 1a) whereas Ph. zapotecus has short and rounded posterolateral angles (Fig. 1c). The basal border of the pygidium in Ph. dionysius forms a small indentation at its middle (Fig. 2b), whereas it runs completely straight in Ph. zapotecus (Fig. 2c). Additionally, the apex of the parameres of Ph. dionysius sp. n. is more projected (Fig. 3b, d), than that from Ph. zapotecus (Fig. 3c, e). Moreover, the middle sinuation of the parameres in lateral view is much more pronounced in Ph. dionysius sp. n. (Fig. 3b) than in Ph. zapotecus (Fig. 3c).
Chorological affinities. The known distribution of Ph. dionysius sp. n. is relatively near to its closest relative, Ph. zapotecus, 90 km distance in a straight line, which is distributed in dry pine-oak and pine-oak-juniper forests on the internal slope of the Sierra Sur (Sierra de Tlaxiaco), going from 1850 m to 2150 m altitude. Interestingly, attempts at trying to collect Ph. zapotecus in the environs of San José del Pacífico in the Sierra Sur (Sierra de Miahuatlán) with fungi-baited traps did not produce any results.

Phanaeus (Notiophanaeus) malyi Arnaud, 2002
Phanaeus pyrois malyi: Arnaud (2002: 96-97). Phanaeus malyi Arnaud, Solís & Kohlmann (2012: 10). Phanaeus pyrois Bates, Edmonds & Zídek (2012: 1, 3, 13, 52-53, 57) Remarks. Solís and Kohlmann (2012) elevated Phanaeus pyrois malyi Arnaud to species status based on the results of mitochondrial DNA studies of Costa Rican specimens that clearly separated Caribbean from Pacific populations at a genetic distance concordant with species level (average Kimura-2-parameter [K2P] = 3.8 %, Solís and Kohlmann 2012). Moreover, Ph. malyi has short and rounded postero-lateral pronotal angles ( Fig. 8a-b), whereas they are long and slender in Ph. pyrois (Fig. 8c-d). Subsequently, Edmonds and Zídek (2012) synonymised Ph. pyrois malyi with Ph. pyrois, pending a comparative analysis of black varieties of Panamanian and South American populations. Phanaeus malyi was compared with material from J. Blackaller's private collection and photos sent by François Génier from the Canadian Museum of Nature (Suppl. material 1: Figure S1) of black specimens of Ph. pyrois from Cerro Campana, Province of Panama, Panama (formerly treated as Ph. pyrois olsoufieffi Balthasar) and Playa de Oro, Esmeralda, Ecuador (formerly treated as Ph. pyrois funereus Balthasar). Ph. malyi and Ph. p. olsoufieffi are different taxa. While Ph. malyi has short and rounded posterolateral angles of the pronotum, similar to Ph. endymion, Ph. p. olsoufieffi has the typical pyrois long and slender posterolateral angles of the pronotum. Ph. p. olsoufieffi also exists in Costa Rica, distributed along the foot of the mountains and not in the lowlands like the typical metallic-red pyrois. A more detailed analysis would be required in order to determine if it should again be resurrected into subspecific status. In relation to Ph. p. funereus, these specimens have an intermediate form of the posterolateral angles of the pronotum between malyi and olsoufieffi. Moreover funereus is distinctly dull black whereas malyi is a shiny black. A more detailed analysis is required for resurrecting funereus, but it is a different taxon from malyi.
We therefore revalidate Ph. malyi to full species status based on the previously mentioned mitochondrial DNA analysis (Solís and Kohlmann 2012); the characteristic ecology and distribution of this species in tropical rainforest along the southern Pacific coast of Costa Rica, probably extending to the northern Panamanian Pacific coast (Solís and Kohlmann 2012); and the different shapes of the pronota (Suppl. material 1: Figure S1).
Phanaeus malyi follows a characteristic distribution pattern described by Kohlmann and Wilkinson (2007) and by Solís and Kohlmann (2012), where many species living in the Costa Rican southern Pacific tropical forest are sister species of taxa found along the tropical forest of the Caribbean coast. This vicariant speciation pattern (apparently mediated by the emergence of the Talamanca Cordillera that has divided a formerly continuous tropical forest stretching from the Caribbean to the Pacific) has been observed to occur in species of dung beetles, fishes, amphibians, reptiles, and birds (Kohlmann and Wilkinson 2007).

Remarks.
Recently this species has been described as found in Guerrero. Here we complement its poorly known distribution with new localities from the Blackaller private collection: Guerrero (38 specimens   report this species seemingly copro-necrophagous. It is reported here to be also mycetophagous and living not only in pine-oak forest, but also in the cloud forest.

Phanaeus (Notiophanaeus) halffterorum Edmonds, 1979
Phanaeus halffterorum: Edmonds (1979: 99;partim), Halffter and Edmonds (1982: 88-89), Anduaga and Halffter (1991: 157), Delgado-Castillo et al. (1993: 125), Edmonds (1994: 39-43, 101), Anduaga (2000: 125, 130), Arnaud (2002: 95-96), Edmonds (2003: 61, 65), Edmonds (2006: 31, 34, 36), Ceballos et al. (2009: 397), Edmonds and Zídek (2012: 3, 5, 12, 52, 54), Deloya et al. (2014: 77), Moctezuma and Halffter (2017: 52, 54-55), Moctezuma et al. (2017: 115), Lizardo et al. (in press Major males with long and slender posterolateral angles of pronotum (Fig. 1a); pygidium has its basal border forming a small indentation at its middle (Fig. 2b); aedeagus and lamella copulatrix as in Fig. 3a Elytral interstriae distinctly flattened and uniformly dull (more convex and shiny in some Central American populations); striae not strongly impressed basally, anterior ends in most specimens bearing deep punctures rather than large fossae. Male: Pronotal disk dull, velvety smooth medially, finely asperate, brighter laterally. Female: Pronotum evenly convex, lacking anteromedial concavity even in largest specimens, bearing three round, smooth tubercles in transverse line near anterior margin. Head and pronotum highly shiny metallic red or green to nearly completely dull black with metallic red restricted to ridges and isolated areas on anterior part of pronotum; elytra dull to lightly shiny black; pygidium usually metallic red medially, green peripherally, in some completely red or green. Remarks. This species has been recorded only once in the literature, more than 40 years ago, when Howden described it in 1974 in the Sierra Sur (Sierra de Miahuatlán). Using the only two known distribution records reported by Howden (1974), Trotta-Moreu et al. (2008) performed an analysis of the known and potential distribution of G. pecki in the state of Oaxaca using the MaxEnt prediction model (Phillips et al. 2006  Historical. Oaxaca. Duraznal, 17°1'55"N, 96°10'12"W, 1820 m. Notes. So far, this species has only been collected in Oaxaca and seems to be restricted to the Sierra Norte (Sierra de Ixtlán, Sierra Mazateca, and Nudo del Zempoaltépetl) in the northern part of the state. Howden (1964) cites only two historical localities for the state taken from Bates' (1886-1890) work, Duraznal, Oaxaca, and (La) Parada. The latter we have been unable to locate. Martínez and Suárez (2006) subsequently cited the locality of "El Llano de las Flores" for this species. Trotta-Moreu et al. (2008) published a study on the known and potential distribution of Geotrupidae in Mexico. Using the GEOMEX database, they predicted a potential distribution based on six localities (which are not cited in their text) and calculate mean values of temperature (15.4 ±3 °C), precipitation (1357 ±936 mm) and altitude (1793 m), as well as a potential distribution area (3352 km 2 ) for this species in the state of Oaxaca. They ambiguously cite that this species is distributed in the Sierra Madre del Sur, not specifying that actually it is distributed in the Sierra Norte. A recent book on the biodiversity of the Sierra Madre del Sur by Luna-Vega et al. (2016) implies that all mountains in Oaxaca belong to the same biogeographic region called Sierra Madre del Sur. This is not a correct biogeographic regionalisation, as the results of this present paper and another (Kohlmann in press) show great differences in the dung beetle fauna of the Sierra Norte and the Sierra Sur. The Sierra Norte shows considerable influence from elements stemming from the Sierra Madre Oriental and the Trans Mexican Volcanic Belt (Kohlmann in press). Finally, Alvarado et al. (2013) cite three georeferenced localities at La Chinantla. With the present collecting localities, we can recalculate and correct the Trotta-Moreu et al. (2008) too low mean value for altitude for this species.
Taxonomic relationships. Geotrupes pecki and G. viridiobscurus have very similar aedeagi with only small differences on the parameres, as can be observed from the photographs of their genitalia (Fig. 10). They also present small differences in the elytral striae, being more or less punctate or crenulated as well as having a more or less punctate pronotum, over the whole surface (pecki) or concentrated at the pronotal sides (viridiobscurus). Howden (1974) has indicated that G. pecki has a heavily punctate pronotum, but as Fig. 9 clearly shows, this is not always the case and the pronotum can resemble the one of G. viridiobscurus. Due to the significant taxonomic similarity discussed above it is here postulated that this species pair are sister species.
Chorological affinities. The known distribution of G. viridiobscurus in the Sierra Norte is relatively adjacent to its closest taxonomic relative, G. pecki, in the Sierra Sur, only 125 km away in a straight line (Fig. 6). Both species are distributed in pine and pine-oak forests; G. viridiobscurus also has been collected in oak and pine/alder forest. G. pecki has been collected from 2400 m to 3000 m altitude whereas G. viridiobscurus has been collected from 2000 m to 3200 m altitude (the Duraznal historical record with 1800 m altitude does not represent a precise locality). The Trotta-Moreu et al. (2008) mean altitude calculations of 1800 m for both species may be incorrect, and therefore it is probable that their calculations of the other climatic variables (temperature and precipitation), and distribution predictions, also would be erroneous.
Key to the Geotrupes of Oaxaca (modified from Howden 1964Howden , 1974 Edmonds (1994) comments that the endymion group of the subgenus Notiophanaeus represents a recent migration into Middle America and spread in response to the northward expansion of Neotropical forests during the Pleistocene. At the time of his first study (Edmonds 1994) only two species (endymion Harold and halffterorum Edmonds) of this group were known to exist in Mexico. At present we know that there are actually seven species (bravoensis, dionysius sp. n., endymion, halffterorum, huichol, zapotecus, zoque) belonging to the endymion group distributed in Mexico (out of the ten [arletteae, malyi, pyrois] that comprise the whole group), which prompts us to reassess the probable time of entry of this group into Mexico. These species inhabit a plethora of ecosystems, such as dry oak scrub-forest, dry pine-oak and dry pine-oak-juniper forest, cloud forest, and wet pine-oak forest. In the case of Oaxaca, Cevallos-Ferriz and Ramírez (2004) Caballero et al. 2011), as the small morphological differences between these two species pairs suggest. Caballero et al. (2011) indicate in a very interesting study that as the LGM temperatures descended to between 6 and 8 °C, the Permanent Snow Line (PSL) descended 1000 m to 3940 m in the Iztaccíhuatl volcano and 3400 m in the Tancítaro and 3650 m in the Cofre de Perote volcanoes. Based on the results, this same study (Caballero et al. 2011) proposes that the alpine meadow/Pinus border descended by ~1000 m along with distributional and compositional vegetation changes. At the beginning of the Terminal Glacial (15 ky to12 ky cal BP), all the glaciers retreated slowly and intermittently, then markedly after 14 ky cal BP, creating drier conditions in the internal basins of the centre of Mexico. At present, the upper limit of the Pinus forest in the Iztaccíhuatl volcano is located at 4020 m (Caballero et al. 2011). Pinus is established in Mexico at broad temperature (15 °C±10 °C) and precipitation (800 ±150 mm) ranks (Velázquez et al. 2000). Caballero et al. (2011) indicate the existence of a precipitation gradient going from both coasts in Mexico to the inland; therefore, glaciers in the inland (Iztaccíhuatl) did not descend as in mountains with a greater maritime influence (Tancítaro and Cofre de Perote), indicating the existence of less humid environments in the inland regions of central Mexico. An intense deglaciation started in the mountains of central Mexico between 15ky and 14 ky cal BP and continued to about 12. 5 ky cal BP (Caballero et al. 2011). Caballero et al. (2011) indicate that during the LGM, the Pinus upper limit in the mountains of central Mexico descended from 4020 m to ~3000 m and the upper limit of the Quercus forest from 3050 m to ~2150 m, based on a 6°C drop in temperature and a descent of ~1000 m of the PSL.

Discussion
In relation to this lowering of the glaciers during the LGM, Metcalfe (2006) suggests that in central Mexico glaciers and alpine grasslands expanded and there were extensive forests of pine, oak, spruce, and alder. This forest expansion during the LGM allows us to propose a scenario where the mother species of Phanaeus dionysius-zapotecus and of Geotrupes pecki-viridiobscurus occupied a continuous forest condition in the central highlands of Oaxaca, framed by the borders of the Sierra Norte to the North and the Sierra Sur to the South (Fig. 6). Both species pairs are related very closely taxonomically, as suggested in the above-mentioned taxonomic treatment, and also are geographical neighbours. A similar scenario could also be proposed for the mother species of Phanaeus bravoensis-halffterorum (another very closely related species pair); in this case framed by the Mexican Transvolcanic Belt to the North and the Sierra Madre del Sur to the South in the State of Mexico-Guerrero area. The continuity of these forests could have been broken up after the LGM, when glacier lower limits started receding to their present higher elevations following the concomitant increase in temperatures and temperate Nearctic forest movement into higher altitudes. This situation possibly could have initiated a vicariant speciation process as suggested by the small taxonomic differences reported for these species pairs in the present study. This mechanism would be following the sky-islands dynamic (Knowles 2000) where species persist in high-elevation refugia during the interglacials, having a higher probability of diverging, enhanced by reduced gene flow and contrariwise presenting genetic admixture at lower elevations during glacial periods. This proposal is a novel approach for studying and understanding speciation patterns for closely related mountain species in southern Mexico. This study is intended as a taxonomic reference for a subsequent modelling analysis of climate and vegetation shifts in the highlands of Oaxaca for the aforementioned Geotrupes and Phanaeus species.
from Soria, Guanajuato. Jane Segleau, Jane Yeomans, Nichole McVeigh, and Kent McLeod checked the English language. Two anonymous reviewers helped us improve the manuscript. Last but not least, Carlos Granados from EARTH's library greatly assisted in finding bibliography. The first author (BK) thanks EARTH University for the sabbatical leave that allowed him to do the research for this paper.