0. AA vein: (0) with a remnant cross vein and appearing as a small “T” proximally; (1) lacking remnant cross vein. CI = 0.16, RI = 0.16.

1. MP vein length compared to RP: (0) normal (long), approximately 30% or more the length of the RP; (1) intermediate, approximately 20–30% the RP length; (2) short, only near margin and not extending backwards (less than 20% the RP length); (3) MP vein absent. CI = 0.30, RI = 0.58.

2. Pigment patch near apical 1/3: (0) absent; (1) present and distal edge perpendicular to longitudinal axis of the body (Note that the patch is sometimes poorly defined; (2) present and distal edge oblique. CI = 0.28, RI = 0.0.

3. Anal region: (0) large and well developed (noticeably expanded); (1) lacking this shape. CI = 0.10, RI = 0.43.

4. Anal region posterior notch: (0) present; (1) absent. CI = 0.11, RI = 0.27.

5. Jugal and AP vein: (0) both present; (1) AP only; (2) both absent; (3) jugal only. CI = 0.17, RI = 0.22.

6. MP vein near posterior margin of wing: (0) straight to very slightly curved; (1) distinctly, strongly curved; (2) does not reach margin. CI = 0.20, RI = 0.0.

7. RA vein extension along apical margin: (0) slight, not reaching the margin; (1) great, reaching and running parallel to margin for part of its length. CI = 0.25, RI = 0.75.

8. AA vein posteriorly (near wing base): bridge vein extension towards CuA vein: (0) forming cell that is completely closed; (1) forming cell partially closed and extending on both sides of AA and AA vein with small break; (2) forming cell partially closed and with extension toward costal margin but large break on AA vein; (3) cell broadly open, bridge vein absent; (4) forming cell slightly closed with bend towards costal margin; (5) forming cell slightly closed and proximal end of bridge vein detached from AA vein; (6) cell bridge vein complete but cell partially open due to break in AA vein; (7) cell absent (lacking bridge vein) and AA vein smoothly curved; (8) cell partially closed with bridge vein not reaching CuA vein; (9) cell absent (bridge vein absent) with vein slightly sigmodal in shape. CI = 0.39, RI = 0.36.

9. Apical margin: (0) broadly emarginate and not projecting medially; (1) narrowly, smoothly emarginate medially; (2) projecting medially; (3) approximately truncate. CI = 0.25, RI = 0.40.

10. Pigmentation: (0) extremely heavy and dark colored; (1) heavy and dark color absent. CI = 0.50, RI = 0.0.

11. Lateral comb (chaetopariae) apical extension: (0) to or near lateral margin; (1) to front margin first and curving towards lateral margin; (2) distant from both front and lateral margin. CI = 0.40, RI = 0.0.

12. Lateral combs (chaetopariae): (0) relatively parallel and straight but curving slightly inwards apically; (1) relatively parallel and straight throughout; (2) curved throughout. CI = 0.50, RI = 0.66.

13. Anterior median ventral process base (crepis): (0) long (length much longer than width; (1) short (length approximately equal to width). CI = 0.50, RI = 0.50.

14. Anterior median ventral process basally (crepis): (0) distinct and triangular (sometimes elongately); (1) indistinct. CI = 0.50, RI = 0.50.

15. Posterior median tormal process (epitorma): (0) smoothly, broadly tapered throughout: (1) smoothly narrowly tapered; (2) more irregularly shaped. CI = 0.18, RI = 0.10.

16. Lateral tormal processes (apotorma): (0) without any obvious circular-shaped muscle attachment point proximally; (1) with attachment point(s). CI = 0.10, RI = 0.0.

17. Dark pattern on anterior half of ventral surface: (0) broadly triangular without any emargination on either side of peak; (1) moderately broadly triangular and with an emargination on either side of the peak; (2) triangular but with parallel sides at narrow base; (3) broadly rounded. CI = 0.60, RI = 0.50.

18. Cavity on dorsal side of proximal portion of epipharynx: (0) broadly emarginate; (1) flat to slightly rounded; (2) narrowly, sharply emarginate; (3) narrowly, gradually emarginate; (4) moderately, sharply emarginate; (5) small shallow emargination on either side of middle (two). CI = 0.27, RI = 0.27.

19. Base of anterior median process in dorsal view (crepis): (0) lateral projections visible; (1) projections lacking. CI = 0.11, RI = 0.38.

20. Coarse setae on ventral surface between combs (prophobae): (0) at middle only; (1) relatively evenly scattered throughout. CI = 0.50, RI = 0.75.

21. Third palpomere: (0) indistinctly and appearing as two, with the third very reduced in size; (1) distinctly as three, third only slightly reduced; (2)third absent; (3) third subequal to second. CI = 0.37, RI = 0.28. Note that in state “0,” the third palpomere is very tiny and difficult to observe. It appears though to be a tiny reduced palpomere.

22. Second palpomere shape: (0) distinctly rounded to transverse; (1) elongate; (2) triangular. CI = 0.40, RI = 0.0.

23. Mentum apical edge: (0) shallow V-shaped notch; (1) U-shaped notch; (2) smoothly rounded notch. CI = 0.40, RI = 0.40.

24. Mentum shape: (0) distinctly transverse; (1) length approximately equal to width. CI = 0.10, RI = 0.52.

25. Mentum oblique lateral edge: (0) no notch, steeply angled; (1) no notch, moderately angled; (2) notch present (slight to distinct). CI = 0.18, RI = 0.52.

26. Shape of anterior declivous region of mentum: (0) narrow and parallel sided; (1) elongate triangular shape with acutely pointed apex; (2) broad and parallel sided; (3) broad triangular shape with a broadly rounded apex. CI = 0.75, RI = 0.75.

27. Anterior declivous region of mentum apex: (0) with a dark lip; (1) dark lip absent. CI = 0.33, RI = 0.0.

28. Anterior declivous region of mentum: (0) projecting anteriorly with ventral surface horizontal; (1) or not. CI = 0.12, RI = 0.12.

29. Paraglossal strut laterally at distal apex: (0) expanded and bifurcate; (1) expanded but not bifurcate; (2) approximately parallel sided; (3) sigmoidal; (4) abruptly expanded near apex. CI = 0.22, RI = 0.39.

30. Palpomere strut apex adjacent to paraglossal strut: (0) expanded; (1) unexpanded. CI = 0.50, RI = 0.0.

31. Glossal lobe; (0) large and emarginate; (1) large and narrowly rounded; (2) large and broadly rounded; (3) large and triangular; (4) small and rounded; (5) absent. CI = 0.33, RI = 0.37.

32. Internal apodeme (V or Y shaped usually with one arm relatively short) adjacent to paraglossal apodeme in dorsal view, from anterior to posterior, appearing: (0) Y-shaped; (1) V-shaped; (2) straight, no arm present. CI = 0.25, RI = 0.62.

33. Internal apodeme (as in #32): (0) obliquely angled outwards; (1) longitudinally aligned; (2) obliquely angled inwards. CI = 0.33, RI = 0.66.

34. Internal apodeme, proximal apex: (0) enlarged; (1) not enlarged. CI = 0.12, RI = 0.12.

35. Most proximal apodeme, length from apex to transverse bridge: (0) long; (1) normal (approximately the same length as the width of the transverse bridge). CI = 0.50, RI = 0.50.

36. Antennal vessicles: (0) absent (or possibly obscure and difficult to see); (1) present and easily visible. CI = 0.33, RI = 0.66.

37. Proximal antennomere of the apical club viewed laterally: (0) about 50% longer than wide; (1) approximately 4 to 5 times as long as wide; (2) approximately two times as long as wide. CI = 0.28, RI = 0.66.

38. Apical flanges: (0) trilobed; (1) or not. CI = 0.50, RI = 0.50.

39. Apodeme: (0) with a large flattened single flange (with one lateral edge slightly expanded perpendicularly; (1) large flattened single flange absent. CI = 0.50, RI = 0.66.

40. Apodeme: (0) with a complete oblique suture/carina; (1) with an incomplete oblique suture/carina; (2) suture/carina absent. CI = 0.14, RI = 0.53.

41. Apodeme pocket: (0) deep and elongate; (1) no deep pocket, slight flanges at most. CI = 0.25, RI = 0.14.

42. Apodeme apex: (0) with a short second flange; (1) short second flange absent. CI = 0.50, RI = 0.0.

43. Apodeme shape: (0) broad, smoothly rounded; (1) lacking a broad and smoothly rounded shape. CI = 0.25, RI = 0.66.

44. Clypeal teeth: (0) two and juxtaposed; (1) two and broadly separated; (2) four; (3) single median. CI = 0.36, RI = 0.58.

45. Genal development: (0) strong, acutely angled or less; (1) weak, angle greater than 90 degrees. CI = 0.33, RI = 0.66.

46. Cephalic projections: (0) central horn anteriorly; (1) central horn posteriorly; (2) paired separate projections. CI = 0.30, RI = 0.30.

47. Carinae anterior of eyes: (0) two transverse carinae; (1) one transverse carina; (2) no carinae. CI = 0.22, RI = 0.61. Taxa that have a horn that obscures a carina if it was present were coded with a dash (-).

48. Transverse projection: horn or ridge posterior of eyes: (0) present; (1) absent. CI = 0.50, RI = 0.92. This ridge is located anterior of a second ridge, if present, demarking smooth usually hidden surface posterior from typically exposed part of the head.

49. Eye canthus: (0) not dividing eye; (1) completely dividing eye. CI = 0.20, RI = 0.80.

50. Head posterior margin, dorsally: (0) wide projection with broadly truncate apex; (1) medial moderate triangular projection approximately equal in size to emarginations on either side; (2) medial narrow triangular projection, much narrower than emarginations on either side; (3) medial broad triangular projection, larger than emarginations on either side; (5) medial broad and elongate triangular projection and with emarginations distally with slightly filled in with darken cuticle; (6) medial very broad projection with approximately equal and broad emarginations on either side. CI = 0.42, RI = 0.61.

51. Posterior margin of gula: (0) broadly rounded and strongly projecting; (1) truncate and slightly projecting; (2) shallowly emarginate at middle; (3) narrowly slightly projecting at middle; (4) broadly rounded and weakly projecting; (5) truncate but narrowly emarginate at middle; (6) narrowly strongly projecting at middle. CI = 0.27, RI = 0.30.

52. Setal pattern on anterior margin of gula: (0) transverse band; (1) elongate triangular shape with base positioned distally; (2) approximately equilateral triangle with base positioned proximally; (3) broad triangular shape with base positioned distally. CI = 0.37, RI = 0.0.

53. Paramere apex: (0) elongately pointed; (1) blunt. CI = 0.12, RI = 0.46.

54. Angle of parameres to basal piece where attached: (0) aligned at approximately same angle; (1) approximately 45 degrees; (2) approximately 90 degrees. CI = 0.20, RI = 0.11.

55. Length of parameres to basal piece, excluding narrow projections: (0) nearly equal to length of basal piece; (1) much greater than ½ the length of basal piece; (2) approximately ½ the length of basal piece. CI = 0.66, RI = 0.50.

56. Dorsal plate orientation: (0) vertical; (1) transverse. CI = 1.00, RI = 1.00.

57. Dorsal plate shape: (0) with distinct lateral projections or “tails;” (1) lateral projections lacking. CI = 0.12, RI = 0.58.

58. Transverse ridge defining or separating propygidium from the pygidium: (0) absent; (1) present. CI = 0.25, RI = 0.72.

59. Median groove: (0) distinct and with complete lateral edges; (1) present but indistinct; (2) absent. CI = 0.50, RI = 0.50.

60. Median groove on propygidium: (0) sides parallel; (1) sides converging. CI = 1.00, RI = 1.00.

61. Propygidium shape: (0) even width throughout; (1) longer medially (slightly); (2) narrowed medially. CI = 0.18, RI = 0.10.

62. Propygidium posterior/ventral border medially: (0) angulate emargination anteriorly; (1) rounded posteriorly; (2) none of the previous. CI = 0.40, RI = 0.25.

63. Dorso-lateral edge of propygidium: (0) with a lateral projection creating an overhanging edge; (1) not projecting and angulate, not rounded; (2) none of the previous. CI = 0.22, RI = 0.41.

64. Dark band of pigment at base: (0) present; (1) absent. CI = 0.25, RI = 0.0.

65. Propygidium length compared to width: (0) relatively broad; (1) or narrow. CI = 0.20, RI = 0.0.

66. First ventrite: (0) distinctly visible throughout entire length; (1) reduced to thin “line” near middle. CI = 0.16, RI = 0.50.

67. First ventrite visible portion projecting between metacoxae: (0) distinctly truncate; (1) not distinctly truncate. CI = 0.20, RI = 0.50.

68. Fifth ventrite, males: (0) reduced to a thin line at middle; (1) wider and obvious. CI = 0.50, RI = 0.0. Note: Males were lacking for Liatongus californicus, Anoplophora, Cassolus, Amietina, and Tomogonus and were coded as “?.”

69. Galea shape: (0) longer than wide; (1) wider than long. CI = 0.33, RI = 0.66.

70. Apical palpomere: (0) basal 2/3 distinctly darker than apical 1/3; (1) approximately the same shade throughout. CI = 0.80, RI = 0.42.

71. Dorsal articulatory sclerite: (0) projecting laterally as an acute point; (1) not projecting laterally as an acute point. CI = 0.14, RI = 0.14.

72. Ventral articulatory sclerite (‘V) of the galea (distal half) in dorsal view: (0) abruptly expanded from base to apex on galea brush side at middle; (1) expansion not abrupt but more gradual. CI = 0.50, RI = 0.0.

73. Ventral articulatory sclerite (‘V’) of the galea (proximal half) in dorsal view: (0) parallel sided or nearly so; (1) expanded at or near middle. CI = 0.12, RI = 0.46.

74. Accessory internal sclerite: (0) curved near distal end; (1) curved near proximal end; (2) straight. CI = 0.33, RI = 0.25. Two taxa, Heteroclitopus and Sukelus, were coded as “?” as the structure was too difficult to see clearly.

75. Inner strut of lacinia (nearest to palpifer), distal tip: (0) with notch (sometimes with one tooth greatly reduced); (1) with notch and with one tooth perpendicular to other; (2) notch absent and tapered tip; (3) notch absent and spatulate tip. CI = 0.42, RI = 0.63.

76. Ventral lacinial articulation sclerite (in ventral view): (0) lacking any extension or “tail;” (1) with a short “tail” extending only part of length; (2) with a long “tail” extending most of lacinial length. CI = 0.40, RI = 0.86.

77. Number of “extra” sclerites associated with the galea: (0) one or more; (1) absent. CI = 0.25, RI = 0.62.

78. Large sclerotized plate of the dorsal aticulation sclerite situated between arms of ventral articulatory sclerite (“V) of the galea: (0) present; (1) absent. CI = 0.33, RI = 0.33.

79. Ventral articulatory sclerite (‘V) of the galea, apex of distal arm: (0) truncate; (1) slightly emarginate; (2) bifurcate; (3) pointed or rounded. CI = 0.23, RI = 0.09.

80. Number of teeth on tibia: (0) four; (1) three; (2) two. CI = 0.50, RI = 0.60.

81. Tibia on proximal side opposite outer teeth: (0) with a line of setae associated with a carina; (1) setae absent. CI = 0.11, RI = 0.50.

82. Pronotum anterior opening dorsally: (0) with a slight indentation medially; (1) or more smoothly rounded. CI = 0.12, RI = 0.53.

83. Tibia between second and third teeth, number of microteeth: (0) two or three; (1) one; (2) none. CI = 0.11, RI = 0.15.

84. Male tibia: (0) elongate, narrow, and curved near apex; (1) relatively short, wide and straight throughout. CI = 0.50, RI = 0.0.

85. Male pronotal sculpture: (0) two posterio-lateral ridges; (1) or not. CI = 0.50, RI = 0.50.

86. Male pronotal projections or depressions: (0) present; (1) absent, surface smoothly rounded. CI = 0.12, RI = 0.61.

87. Pronotum with pale colored maculations: (0) present; (1) absent. CI = 0.25, RI = 0.0.

88. Pronotal surface: (0) glabrous; (1) setose, although sometimes only at or near edges. CI = 0.14, RI = 0.40.

89. Pronotum posterior margin dorsally: (0) with a slight but distinct posteriorly directed point or angulate edge; (1) edge rounded. CI = 0.14, RI = 0.33.

90. Apex, appearing (0) sharply acute; (1) narrowly rounded; (2) broadly rounded. CI = 0.66, RI = 0.87.

91. Outer lateral edge: (0) bulbous/expanded laterally; (1) no bulbous expansion laterally. CI = 0.50, RI = 0.0.

92. Apical extension of pigmented mesal comb compared to sclerotized (darkened) area near opposite lateral edge: (0) longer; (1) absent (not visible); (2) approximately the same length or slightly shorter; (3) approximately ¾ the length of the sclerotized area. CI = 0.60, RI = 0.66.

93. Outer lateral edge and part of ventral surface: (0) with bright chestnut red colored cuticle; (1) red colored cuticle absent. CI = 1.00, RI = 1.00.

94. Elytral striae: (0) composed of a single line; (1) composed of a three lines creating two rows (i.e., with a thin line separating two adjacent rows). CI = 0.33, RI = 0.33

95. Pigment patches: (0) dark elongate areas on some intervals adjacent to pale cuticle; (1) dark areas lacking. CI = 0.20, RI = 0.0.

96. First, third and fifth intervals: (0) slightly raised above surrounding intervals; (1) nearly the same height as the others. CI = 0.25, RI = 0.25.

97. Humeral angle in dorsal view: (0) with a lateral bulge; (1) bulge absent. CI = 0.50, RI = 0.50.

98. Epipleura: (0) double; (1) single; (2) indistinct. CI = 0.66, RI = 0.0.

Note in Tomogonus there is no distinct epipleura but the elytra are smoothly rounded to the lateral edge.

99. Eighth interval: (0) extending only approximately ½ the length of the elytron; (1) or not. CI = 1.00, RI = 1.00.

Note that in Scaptocnemis and Tiniocellus, the eight interval appears to be obsolete beyond the basal half.

100. Seventh stria: (0) strongly and distinctly curved inwards towards 6^th stria near the elytral base; (1) slightly curved at most. CI = 0.16, RI = 0.54.

101. Shape at anterior margin: (0) rounded without or only a slight lateral protrusion; (1) or not. CI = 0.50, RI = 0.50.

102. Lateral margin near apex: (0) with a distinct beak-like tapered straight point; (1) tapered straight point absent. CI = 0.33, RI = 0.71.

103. Apex: (0) with a distinct darkened hook projecting ventrally; (1) hook projecting ventrally absent. CI = 0.11, RI = 0.50.

104. Apex with a lateral pocket or cavity: (0) present; (1) absent. CI = 1.00, RI = 1.00.

105. Apex, including muscle apodeme: (0) broadly rounded; (1) or not. CI = 0.16, RI = 0.44.

106. Scutellum apex: (0) elongate and parallel sided, tongue-shaped; (1) triangular shaped. CI = 0.25, RI = 0.78.

107. Prescutum anteriorly: transverse plate approximately ventrally directed: (0) narrow, parallel sided, needle-shaped; (1) angularly emarginate to various depths; (2) narrowly rounded; (3) short small projection; (4) flat, no projection; (5) broadly rounded. CI = 0.55, RI = 0.42.

108. Prescutum anteriorly transverse plate approximately ventrally directed: (0) with distal tips curved out laterally; (1) tips parallel or absent. CI = 0.33, RI = 0.0.

109. Prescutum-scutum junction when viewed in a horizontal position: (0) sharply emarginate; (1) broadly emarginate; (2) broadly rounded; (3) narrowly rounded at middle; (4) straight. CI = 0.33, RI = 0.33.

110. Scutum- scutellum, each sclerite: (0) on two different planes or levels (with a strong declivity between them); (1) on a single plane or level (declivity absent or weak). CI = 0.50, RI = 0.50.

111. Anterior margin of scutum: (0) with a pale colored, transverse region of cuticle; (1) pale colored transverse region absent. CI = 0.14, RI = 0.25.

112. Prescutum: prephragma (longitudinally directed median apodeme with scutum in a horizontal position: (0) with a transverse suture; (1) transverse suture absent. CI = 0.50, RI = 0.50.

113. Scutum: (0) distinctly transverse; (1) not distinctly transverse in shape. CI = 0.16, RI = 0.37.

114. Scutum lateral projection (0) extended only slightly and truncate; (1) distinctly extended and with a blunt apex; (2) extended and with a pointed apex; (3) projection absent. CI = 0.37, RI = 0.77.

115. Scutum lateral edge: (0) with a rounded or somewhat rounded lateral pocket; (1) pocket absent. CI = 0.20, RI = 0.76.

116. Prescutum: prephragma in ventral view: (0) recurve absent (and with longitudinal axis): (1) recurve present; (2) recurve absent (longitudinal axis absent to support recurve). CI = 0.40, RI = 0.78.

117. Scutellum ventral surface at apex: (0) with carina extending anteriorly through V-shaped sclerite; (1) carina absent. CI = 1.00, RI = 1.00.

118. Scutellum in ventral view: (0) triangular shaped and much less than ½ the width of the scutum; (1) or not. CI = 0.50, RI = 0.0.

119. Scutellum posterior margin: (0) no margin or edge; (1) slight margin or remnant edge; (2) short, round triangular projection (converging sides); (3) long triangular projection (parallel or subparallel sides for part of length); (4) short projection with truncate apex, gradually converging sides; (5) short projection with truncate apex, rapidly converging sides; (6) short projection, gradually converging sides with emarginate apex; (7) short triangular projection (converging sides) but truncate apex; (8) broad, subparallel projection with truncate apex; (9) short, sharp triangular projection (converging sides). CI = 0.64, RI = 0.80.

120. Scutellum anterior margin: (0) straight to very slightly projecting; (1) moderately projecting and broadly rounded; (2) sinuate; (3) truncate (rounded at anterio-lateral edges); (4) slightly emarginate; (5) strongly projecting and broadly rounded. CI = 0.55, RI = 0.0.

121. Alar ridge sides on scutellum excluding anterior projection towards posterior: (0) narrow in width and sides slightly converging basally; (1) wide and sides strongly converging basally; (2) wide and sides slightly converging basally. CI = 0.40, RI = 0.76.

122. Mediophragma, ventral margin cleft shape: (0) deeply emarginate (sides subparallel); (1) shallowly emarginate (sides obliquely angled). CI = 0.50, RI = 0.0.

123. Mediophragma: (0) with a distinct longitudinal ridge; (1) with a partial ridge (remnant easily visible); (2) ridge absent. CI = 0.50, RI = 0.86.

124. Mesosternum: (0) with a transverse and longitudinal low ridge forming a cross shape of various forms; (1) with a U shaped depression intersected by a longitudinal ridge; (2) with a transverse ridge with a very slight medial expansion anteriorly; (3) transverse ridge medially narrowly smoothly expanded ventrally; (4) with a longitudinal ridge expanded anteriorly; (5) with a transverse ridge expanded slightly anteriorly at middle forming a truncate anterior margin; (6) with a transverse ridge (in Amietina extending anteriorly gradually towards middle). CI = 0.66, RI = 0.57.

125. Meso- and metatibia: (0) narrow and only slightly expanded apically; (1) broad and greatly expanded apically. CI = 0.25, RI = 0.25.

126. Metatarsi: length of first metatarsomere compared to second: (0) first nearly 2X or more the length of second; (1) first only slightly longer than second. CI = 0.33, RI = 0.50.

127. First metatarsomere: (0) lengthened and greatly expanded laterally; (1) neither lengthened and expanded laterally. CI = 0.50, RI = 0.50.

128. First metatarsomere: (0) distinctly fringed on both lateral edges with dense combs of setae; (1) distinct setal fringes lacking. CI = 0.33, RI = 0.0.

129. Metacoxal separation at middle: (0) relatively large, coxae distinctly separated throughout and ventrite projection anteriorly at middle separating coxae truncate at apex; (1) relatively narrow, coxae not distinctly separated anteriorly, in contact or nearly so, ventrite projection pointed at apex. CI = 0.50, RI = 0.85. Note that this character appears to be quite variable within the Eurysternini.

130. Metatibia: (0) elongate and parallel sided at middle ½; (1) or not. CI = 0.25, RI = 0.57.

131. Metasternum between mesocoxae: (0) diverging anteriorly; (1) approximately parallel sided; (2) converging anteriorly. CI = 0.22, RI = 0.58.

132. Metasternum posteriorly: (0) with a shallow depression; (1) depression absent. CI = 0.33, RI = 0.0.

133. Metendosternite basal stalk: (0) sides diverging throughout length; (1) sides parallel (sometimes slightly sigmoidal) to slightly converging . CI = 0.33, RI = 0.33.

In all analyses, the Helictopleurina and Drepanocerina are monophyletic (supported by four and six controverted synapomorphies respectively) while the Oniticellina is paraphyletic. The sister relationships of Tiniocellus + Scaptocnemis, Tragiscus + Oniticellus, and Anoplodrepanus + Euoniticellus are also invariant in all analyses. Attavicinus monstrosus (Bates), formerly placed in Liatongus (see Philips and Bell 2008), does not share common ancestry with the other Liatongus taxa. The clade consisting of Attavicinus + Paroniticellus is maintained except in the most extreme weighting (K = 1 (and 2, not illustrated); Fig. 3). In those two topologies, Paroniticellus is shifted to a more derived position in the tree and Attavicinus is a basal lineage and sister to all remaining Oniticellina + Drepanocerina. Also, the clade composed of Liatongus militaris as sister to Tragiscus + Oniticellus is shifted to an apical origin. The remaining oniticellines show only minor differences in arrangement in the trees with K values < 6 compared to that seen in with K values of 8–10.

This tribe is a well-defined monophyletic group based on both Bayesian and parsimony analyses, and using either unweighted or weighted data. In the unweighted topology, five states (one uncontroverted) support monophyly. The relationships found with K values = 8–10 are identical (Fig. 1B) and differ from those found at higher K values of 11–30 (Fig. 5) in that the basally originating Phalops and Digitonthophagus are sister taxa in the former instead of successive branches in the latter. There are also minor differences of the position of some taxa in the middle of the tree (from Onthophagus hecate through to Milichus). Trees based on stronger weighting with K = 1–7 values (see Fig. 3) place Digitonthophagus as basal and, in trees from K = 2–7, Onthophagus capella + Diastellopalpus as the next most apical clade. In the K = 4–7 trees, a third basal clade consisting of Proagoderus + Phalops is seen.

The trees found with the most extreme weighting (K values of 1–3) compared to all others are the most different (Fig. 3); the topology found with K = 2 (not illustrated) contains elements of both. Changes include the shifting to a more apical origin of both Proagoderus and Phalops.

The probable symphile taxa (from Amietina to the tree apex) with the most derived features have strong support for monophyly; all included nodes are supported by a minimum of one uncontroverted and two controverted states. These lineages are not dung feeders but necrophilous (Amietina), myrmecophilous (Haroldius, Megaponerophilus, Alloscelus, and likely Eusaproecius and Pseudosaproecius) or termitophilous (Heteroclitopus, Pinacotasus, Stiptopodius, and probably Sukelus). Inquiliny most parsimoniously evolved only once in the common ancestor of Haroldius through to Sukelus. Support for this clade is based on five character states, including one uncontroverted. The Alloscelina (Alloscelus, Haroldius, and Megaponerophilus) is found within this larger clade but there is no support for its monophyly in any of the tree searches performed.

Traditionally the Oniticellini are divided into three subtribes, the Drepanocerina, Helictopleurina, and the Oniticellina (e.g., Halffter and Matthews 1966). In this study the Drepanocerina and Helictopleurina are monophyletic. Recognition of the Helictopleurina as a tribe would eliminate possible paraphyly in the Oniticellini. This was done for perhaps the first time in Orsini et al. (2007), although their study provides little evidence for doing so with a single outgroup and only Madagascar species represented.

The Drepanocerina currently contain 12 genera but until recently included only two, Cyptochirus and Drepanocerus. Even the genus Cyptochirus was previously considered a synonym of Drepanocerus (e.g., Arrow 1931, Halffter and Matthews 1966). Simonis and Zunino (1980) validated the genus, which at that time included several Oriental species now placed in Sinodrepanus (Simonis 1985a). Simonis (1985b) later hypothesized Drepanoplatynus and Sinodrepanus as sister genera and this clade in turn sister to Drepanocerus, relationships not supported in Roggero et al. (2015). New genera in this subtribe were recently created (Krikken 2009, Barbero et al. 2009a,b, and Roggero et al. 2015) including Afrodrepanus Krikken, Clypeodrepanus Krikken, Drepanellus Barbero, Palestrini and Roggero (now a junior synonym of Latodrepanus), Eodrepanus Barbero, Palestrini and Roggero, Epidrepanus Roggero, Barbero, and Palestrini, Latodrepanus Krikken, Paraixodina Roggero, Barbero, and Palestrini and Tibiodrepanus Krikken. A recent phylogenetic study of morphology by Roggero et al. (2015) including representatives of all genera has now hypothesized relationships within this subtribe.

The two representatives in this study, Cyptochirus and Afrodrepanus, were found as sister taxa. The Drepanocerina may not be monophyletic based on preliminary studies mentioned but not discussed in Barbero et al. (2009b), perhaps due to the relative heterogeneous morphologies of the genera. Krikken (2009) does give a morphological description of the group but later notes that support for their monophyly is negligible. Roggero et al. (2015) using Anoplodrepanus as a single outgroup gives a relatively weak test for monophyly. But based on the results in this study and the hypothesized close relationship of taxa included within the Drepanocerina, there is support for the monophyly of the clade. Notably, Scaptocnemis and Anoplodrepanus have been placed by some authors in the Drepanocerina previously, but this study supports their placement in the Oniticellina as suggested earlier by Branco (2010) and Krikken (2009). The Drepanocerina as defined herein is therefore restricted to the Afrotropical-Oriental regions.

The Madagascar endemic helictopleurines are composed of only two genera, the speciose Helictopleurus with ~80 species and the monotypic Heterosyphus. The included species of Helictopleurus representing three subgenera are strongly supported as monophyletic. The rare Heterosyphus was not available for study. When Paulian (1975) described this genus he placed it in the Canthonini and discussed its unusual morphology, surmising that it was not closely allied with this clade. But he also stated that this taxon might be related to Helictopleurus based on the form of the head and legs. Similarities were noted with Sisyphus in the form and sculpture of the prothorax and that the presence of a visible scutellum separates it from Onthophagus, although this last character is present to various extent within species of Helictopleurus as well. Halffter and Edmonds (1982) may have been the first to place Heterosyphus in the Helictopleurina. The fairly well supported molecular study of Wirta et al. (2008) shows this taxon to be a derived Helictopleurus. Therefore, Heterosyphus Paulian should be considered a junior synonym of Helictopleurus d’Orbigny so that the species becomes Helictopleurus sicardi (Paulian).

The position of the Helictopleurina in the moderately or heavily weighted topologies supports the recognition of the group as either a monophyletic subtribe or tribe. In contrast, the unweighted or slightly weighted topologies support the creation of a new tribe to avoid a paraphyletic Oniticellini. But molecular evidence from Monaghan et al. (2007) in some of their analyses suggests that the helictopleurines are derived from within the oniticellines. Therefore although the results found in this study support the helictopleurines as either a subtribe or tribe, current molecular evidence suggests that this clade should continue to be recognized as a subtribe.

The subtribe Oniticellina is paraphyletic, although some generic sister relationships are invariant (see in results above). Euoniticellus and Paroniticellus have been classified as subgenera of Oniticellus in the past (Halffter and Matthews 1966) but neither appears phylogenetically close to Oniticellus and their generic status is supported. Interestingly, some evidence exists that the monotypic Paroniticellus and Oniticellus, (perhaps only the subgenus Oniticellus, sensu stricto) are closely related, as larvae of both are characterized by numerous setae on the head capsule (Lumaret 1979). But immatures will first need to be examined over a broader range of species to confirm this potential synapomorphy. The Jamaican endemic Anoplodrepanus is sister to Euoniticellus and also probably a valid genus.

While an African Euoniticellus was included in this study, the single New World species E. cubiensis Laporte from the West Indies (Cuba, Jamaica, Isla de la Juventud, and the Bahamas (Woodruff 1973, M.A. Ivie pers. comm.)) was not. Zunino (1982) states that this species (referred to as Oniticellus cubiensis) “belongs though in a rather isolated position, to the genus Euoniticellus.” A more recent phylogeny of Euoniticellus by Cambefort (1996) solidly supports its placement in this genus.

Several genera not included also deserve brief comments. The monotypic genus Scaptodera was removed from generic synonymy by Paulian (1986), but might be a Liatongus based on its previous placement. Ixodina is often considered a synonym of Drepanocerus (e.g., Davis et al. 2008) but Krikken (2009) and Roggero et al. (2015) list this genus as valid and part of the Drepanocerina. The species now placed in the Southeast Asian Yvescambefortius (Ochi and Kon 1996) was originally placed in Oniticellus by Gillet (1926) and later in Tiniocellus by Janssens (1953). Ochi and Kon (1996) hypothesized a close relationship of their new genus with the east African Scaptocnemis, which in this phylogeny is sister to Tiniocellus. Hence, it may be part of the Scaptocnemis + Tiniocellus clade.

Based on this study, a solution for monophyly within the Oniticellini subtribes outside of the Helictopleurina would be to recognize all genera (including the Drepanocerina) in a single subtribe, or alternatively, the recognition of the Drepanocerina and two or more subtribes. From the Bayesian analysis and the majority of parsimony topologies found (see Figs 1A, 4B), another possibility is a subtribe including Paroniticellus and Attavincinus, while the other would include all remaining genera together with the Drepanocerina. One could also divide the Oniticellina into three or more subtribes while continuing to recognize the Drepanocerina. Based on current evidence of relationships, a possible subtribal classification recognizing hypothesized monophyletic groups is as follows:

Drepanocerina: Afrodrepanus, Clypeodrepanus, Cyptochirus, Drepanocerus, Drepanoplatynus, Eodrepanus, Epidrepanus, Ixodina, Latodrepanus, Paraixodina, Sinodrepanus and Tibiodrepanus.

Helictopleurina: Helictopleurus (with Heterosyphus a junior synonym of Helictopleurus).

Liatongina subtr. n. Philips. Type genus Liatongus Reitter, 1893: Liatongus Reitter (monophyly of the genus needs confirmation).

Diagnosis: This subtribe can be characterized by the following two characters: The mesonotal prescutum anteriorly with the transverse plate approximately ventrally directed is narrowly rounded apically and the posterior median tormal process (epitorma) of the epipharynx is smoothly and broadly tapered throughout its length. Based on a weighted topology (K=3) the three following characters also support this clade: The prothoracic apodeme has an incomplete oblique suture/carina, one or more “extra” internal sclerites are associated with the galea, and the metanotum scutellum posteriorly has a slight margin or remnant edge.

Oniticellina: Anoplodrepanus, Euoniticellus, Nitiocellus, Oniticellus, ScaptocnemisScaptodera, Tiniocellus, Tragiscus and Yvescambefortius.

Attavicina subtr. n. Philips. Type genus Attavincinus Philips & Bell, 2008: Attavicinus and Paroniticellus.

Diagnosis: This subtribe can be characterized by the following combination of characters: Maxilla internal accessory sclerite is curved near the proximal apex, the protibia on the proximal side has a line of setae associated with a carina, males have a longitudinal ridge on the pronotum laterally, and the mesosternal scutum is distinctly transverse in shape.

Note that based on the most heavily weighted parsimony topologies (Piwe K values = 1 and 2), Paroniticellus should be placed within the Oniticellina. Based on the Bayesian topology, Paroniticellus may need to be placed in its own or possibly one of the other subtribes, as its relationship to the other subtribes (excluding the Helictopleurina) is unclear.

This tribe in every analysis was a well supported monophyletic group. Philips et al. (2004b) also support the monophyly of this tribe based on morphological evidence. In contrast, molecular data in Monaghan et al. (2007), Wirta et al. (2008) and Mlambo et al. (2015) generally support a poly- or paraphyletic Onthophagini and indicate that some currently recognized onthophagines may belong elsewhere.

The topologies found herein with extreme weighting K = 1–3 topologies are considered less trustworthy compared to those found with higher K values. This conclusion is also supported by the molecular studies of Monaghan et al. (2007) and Wirta et al. (2008) that do not support either a relatively derived position of Proagoderus and Phalops. Further, Tarasov and Solodovnikov (2011) also found support for a basal origin of Proagoderus and Diastellopalpus as seen in the less strongly weighted topologies. An internal classification of the onthophagines, if desired and based on the evidence herein, would likely necessitate the creation of numerous subtribes to maintain monophyly in classification.

Branco (1989, 1990, 1991, 1992a,b, 1995) placed several of these most derived genera together in a new group based on shared morphological features, such as antennomere number. His Pinacotarsus group includes eight genera (taxa in bold included in this study): Dorbignyolus, Eusaproecius, Heteroclitopus, Krikkenius, Pinacopodius, Sukelus (as Falcidius), Stiptocnemis, and Pinacotarsus. The Stiptopodius group, first defined by Boucomont (1923) and containing some of the genera in the Pinacotarsus group, was redefined by Branco (1995) and includes the genera Neosaproecius, Pseudosaproecius, Stiptopodius, and Stiptotarsus. Of the six genera included in this study, all form a monophyletic clade at the apex of the onthophagines in both the parsimony and Bayesian analyses. Branco hypothesized the Stiptopodius group as less derived than the Pinacotarsus group and this is approximately seen in Figs 1 and 2. But neither group is monophyletic in any of the topologies discovered in this study.

As a very preliminary test of monophyly for Onthophagus, three species were studied, including one Nearctic, an Australian, and one Ethiopian. Also included were four taxa that are usually considered subgenera of Onthophagus (i.e., Digitonthophagus, Diastellopalpus, Proagoderus, and Strandius), as well as Euonthophagus, Hyalonthophagus, and Mimonthophagus. All will be considered Onthophagus, sensu lato in the discussion below.

In this study, Onthophagus, in either the strict or broad sense, is not monophyletic. Onthophagus, sensu stricto, typically appears in three clades. Onthophagus, sensu lato, appear in various topologies in no fewer than four to as many as nine separate lineages. Strandius often appears as part of a clade of Onthophagus while Euonthophagus, Hyalonthophagus and Mimonthophagus are justified as genera evolutionarily distinct from Onthophagus. In regards to Onthophagus, sensu stricto, molecular evidence from Monaghan et al. (2007) also does not support monophyly. Tarasov and Kabakov (2010) discuss the difficulty with placing species within the subgeneric classification. Furthermore, the creation of subgenera based on Asian species that are derived from African lineages will certainly create issues of monophyly for the latter.

Two phylogenetic studies have been done that concentrate mainly on Onthophagus. One on the “Serrophorus” complex (Tarasov and Solodovnikov 2011) used morphological data and included 39 species of Onthophagus. This work also explored the use of male endophallic characters in the group; from 74 parsimony informative characters used, 45 (~60%) were coded from male genitalia. Emlen et al. (2005) used a sample of 48 Onthophagus species, employing up to seven gene sequences. Monophyly of the included taxa is assumed in both studies and it is possible that this is not the case. Regardless, more taxonomically broad studies on this genus and with molecular data are certainly needed to clarify relationships among subgenera and species groups. More genera then are currently recognized will likely be needed to delimit monophyletic clades in this speciose genus.

From various sources (Wikispecies website accessed 3 N. 2015, Zunino 1979, Lumaret and Kim 1989, Tarasov and Solodovnikov 2011, Tagliaferri et al. 2012), the subgenera of Onthophagus that perhaps are most commonly accepted are listed below (taxa in bold included in this study). Excluding five that most dung beetle workers recognize at the generic level (Diastellopalpus, Digitonthophagus, Mimonthophagus, Proagoderus and Strandius), this leaves 25 subgenera as follows: O. (Afrostrandius), O. (Altonthophagus), O. (Amphionthophagus), O. (Bicornonthophagus), O. (Colobonthophagus), O. (Euonthophagus), O. (Exonthophagus), O. (Furconthophagus), O. (Gibbonthophagus), O. (Hyalonthophagus), O. (Indachorius), O. (Macronthophagus), O. (Matashia), O. (Micronthophagus), O. (Onthophagiellus), O. (Onthophagus), O. (Palaeonthophagus), O. (Paraphanaeomorphus), O. (Parascatonomus), O. (Parentius), O. (Pseudophanaeomorphus), O. (Relictonthophagus), O. (Serrophorus), O. (Sinonthophagus) and O. (Sunenaga). Onthophagus (Phanaeomorphus) is another group that could be recognized (Lumaret and Kim 1989), as are O. (Trichonthphagus), O. (Eremonthophagus) and O. (Palaeonthophagus) (Zunino, 1979), and O. (Pseudonthophagus) Kabakov and Shokhin 2014, and see Tagliaferri et al. (2012 for the latest added above) bringing the total number of subgenera to as many as 30 (see Tarasov and Kabakov 2010 for a review of the subgenera).

There are a number of scarabaeine species that historically have been placed in various tribes, reflecting difficulty in their classification. This often included some of the inquilinous scarabaeines with highly modified morphologies, such as flattened leg segments and the presence of trichomes (e.g., Krell and Philips 2010). The formerly recognized Alloscelina (or Alloscelini) was once placed within the Scarabaeini, which included all the “rollers” at that time (Halffter and Matthews 1966). This group included the currently recognized onthophagine genera Alloscelus, Megaponerophilus (a former subgenus of Caccobius), Haroldius (including its junior synonyms Ponerotrogus Silvestri and Afroharoldius Janssens, both synonymized by Paulian (1985), Formicdubius Philips and Scholtz, synonymized by Krell and Philips (2010)) and the genus Freyus Balthasar (now a junior synonym of the dichotomine Paraphytus Harold (Halffter and Matthews 1966)).

Two problematic genera are Haroldius Bocomont and Cassolus Sharp. Hanski and Cambefort (1991) classify Haroldius as a Canthonini, while Philips and Scholtz (2000) placed it within the Onthophagini. Cassolus is listed as a canthonine in Halffter and Matthews (1966), but as an onthophagine in Hanski and Cambefort (1991). In this study, Cassolus and Haroldius appear deep within the onthophagines as sister taxa and supported by 12 character states, including two uncontroverted. Together with their sister clade, they are supported by five characters, including one that is uncontroverted. But they may indeed not be onthophagines. In Bai et al. (2011) using morphological evidence, they found Cassolus + Parachorius positioned outside the onthophagine clade. A recently described species of the Oriental canthonine Parachorius bridges the morphological gap between these two genera, and both form a monophyletic clade according to Tarasov and Keith (2011). At this time, the effects of morphological convergence remain unclear and evidence from molecular data is probably needed to resolve the origins of both genera and their proper tribal classification.

In Philips et al. (2004b), the Onthophagini and Oniticellini shared what was thought to be a fairly strong and intriguing synapomorphy of antennal cavities in the first (proximal) and middle antennomeres of the club. It was also seen in the canthonine Epirinus hilaris Péringuey but is lacking in E. silvestris Cambefort and seven other canthonines examined in Philips et al. (2004b). Interestingly, E. hilaris plus three other Epirinus species may be sister to the sisyphines, onitines, oniticellines and onthophagines, as seen in the molecular based topology of Monaghan et al. (2007). Some onitine taxa later examined, notably Platyonitis smeenkorum Krikken, were also noted to share this feature, while others (e.g. Bubas sp., Heteronitis sp., and Onitis spp.) did not. Interestingly, this feature is also found in the dichotomines Ontherus Erichson, Scatimus and “Trichillum” but not in Scatrichus and “Pedaridium” (F. Génier pers. comm.). Recently, Tarasov and Solodovnikov (2011) discovered that the Onthophagus subgenus Parascatonomus has only a single cavity on the first antennomere of the club. As these cavities are absent in a reasonably broad selection of other dichotomiines and coprines (and hence it is probably not a plesiomorphic condition for dung beetles), this trait appears to have independently evolved numerous times. The function of these antennal cavities is unknown.

Ancestral oniticellines and onthophagines are presumed to have been coprophagous as all of the basal lineages are dung feeders. Tunneling behavior in this clade is very similar amongst taxa and all are classified as Pattern I nesters in Halffter and Edmonds (1982), with a few notable exceptions. An alternative nesting strategy known only within the Oniticellini is dwelling or endocoprid behavior (classified as Pattern VII nesters), which can include nests within dung or nests in pits directly below the dung. Known only in Oniticellus and Tragiscus (Davis, 1989), it may also occur in Paroniticellus although this is currently unclear (see Halffter and Matthews 1966 and Philips and Bell 2008 for comments). Excluding Paroniticellus, this behavior only evolved once in the common ancestor of the sister genera Oniticellus and Tragiscus.

Both tribes are characterized typically and most distinctively by the absence of a brood ball and larvae are supplied with a brood mass that is either modified to some degree or not. Additionally, male-female cooperation is absent or limited and brood care is not known to exist. These are simple behaviors in what one could argue is a derived lineage compared to some of the more ancient origin tunneling clades where complex behaviors evolved. For example, coprines construct brood ovoids, and have extensive male-female cooperation, and brood care. But simple nesting behavior appears to be a very successful strategy evolutionary; the oniticellines and onthophagines are some of the most successful dung beetles in many ecosystems in terms of number of species, and often in terms of abundance and biomass (Davis et al. 2008).

The African endemic Stiptopodius and Pinacotarsus groups of onthophagines form a monophyletic clade as discussed above. Based on the results, all included species are probably either myrmecophilous or termitophilous. As evidenced by the relatively high diversity within this lineage, the association with social insects may be relatively ancient. In contrast, only the New World oniticelline, Attavicinus montrosus, is an inquiline. This species is associated with leaf-cutter ant debris piles and has a very restricted distribution in central Mexico (Philips and Bell 2008). There is no strong evidence for the age of these associations at this time.

Krikken (1982) stated that the onthophagine genera Dorbignyolus, Krikkenius and Pinacopodius are all termitophilous. It is also thought that several others, including Pinacotarsus, Heteroclitopus, Stiptocnemis and Sukelus, are also known or likely termitophilous (Branco 1991, 1992a,b). Directly above this termitophilous clade in a less derived position are two taxa in the genera Eusaproecius and Pseudosaproecius. Pseudosaproecius is considered a generalist in regards to food source by Branco (1995), with two species collected on fish carrion and a third on human dung (and see Cambefort 1984). But he also notes that nine species have been collected only at light. Similarly, only a single record from dung of a Eusaproecius species (of seven described) is known (Branco 1992a). The isolated dung and carrion records are probably not typical, but more likely may be supplemental food sources. A total of 26 specimens of Neosaproecius trituberculatus (Frey) were only collected in flight intercept traps and never at any bait in Ghana (Philips, unpublished data). Hence, it seems probable based on common ancestry, all of the species in this clade are symphiles of some type.

Above these clades (in less derived positions) in successive steps are three additional lineages also associated with social insects. The first, Alloscelus, is thought to be closely related to Pseudosaproecius (due to similar male genitalia and form of sexual dimorphism) but they do not form a monophyletic clade. The second and third, Megaponerophilus and Haroldius + Cassolus, are mainly myrmecophiles, with the exception of Cassolus where only one species of nine is known to be an ant associate (Halffter and Matthews 1966).

In summary, the evidence in this analysis suggests that there may be a single origin of inquiliny in the onthophagines. Several genera with unknown habits are probably associated with either ants (Eusaproecius and Pseudosaproecius) or termites (Sukelus). The other four genera may also be considered as part of this clade although future studies may shift Haroldius and Cassolus out of the onthophagines. The evolution of myrmecophily preadapting the more derived taxa for an association with the Isoptera also seems plausible.

Necrophagy and mycetophagy are rare within the oniticellines and onthophagines and perhaps are recently evolved behaviors with several independent origins. For example, Liatongus rhinocerulus (Morón 1984) and some species of Onthophagus are specialists on basidiomycete mushroom fruiting bodies, although L. rhinocerulus has also been found on carrion and most of the Onthophagus are probably generalist feeders (Halffter and Matthews 1966). Carrion feeding is a specialty of species of Amietina, is known in Caccobius, and in New and Old World species of Onthophagus. There are many taxa associated with feeding on millipede carrion as well, particularly in Africa, perhaps indicating a more ancient origin of this habit. Other alternative feeding habits, such as fruit feeding, may be more recently derived. One should note that the use of any of these foods by adults does not necessarily indicate their use as larval brood food.

The Oniticellini and Onthophagini are postulated to be a relatively modern aged group of dung beetles by Cambefort (1991). Davis et al. (2002) noted that the antiquity of generic distribution patterns lent support for an old age of the entire tribal and subtribal lineage divergence within the Scarabaeinae. Estimates for an origin of the common ancestor of these two tribes are the late Mesozoic (Davis et al. 2008, Philips 2011). More precisely, Mlambo et al. (2015) place the origin at ~early Miocene and the Oniticellini at mid to late Miocene. An Afrotropical origin of the common ancestor of both tribes is most likely, based on this study as well as Philips et al. (2004b) and Monaghan et al. (2007). Based on the results herein, most of the lineages are restricted to the Afrotropical region (Fig. 6). While all of the early branching as well as most of the generic level lineages of onthophagines are restricted to the Afrotropics with dispersals to other regions found in typically more derived lineages, the oniticellines show a pattern of relatively ancient lineage shifts into Madagascar, Palaearctic, and Nearctic regions and, more recently, into the Orient and the Caribbean.

Another general trend of the Oniticellini appears to be successive invasions from the Ethiopian into the Palaearctic through to the Oriental region (including the Lesser Sunda Islands and Sulawesi) in many groups (e.g., Tibiodrepanus (Barbero et al. 2011)). Currently the Drepanocerina have eight genera endemic to Africa while the Oniticellina have two. Hence, 14 generic level clades that evolved in Africa have dispersed to other geographic regions. In the Palearctic and the Caribbean, the Oniticellina have a single endemic genus in each region (Paroniticellus and Anoplodrepanus, respectively), while the Drepanocerina have a single endemic (Sinodrepanus) in Southeast Asia and southern China (Davis et al. 2002). The Oniticellina are known from the Mediterranean region but, in contrast, no Drepanocerina are currently found there, although fossil deposits from England dated at 70–100,000 years BP indicate they were once present (Krikken 2009, Barbero et al. 2009b).

Based on basal branching, the oniticelline ancestors that led to the Madagascan helicopleurines, the Palaearctic Paroniticellus, and the New World Attavicinus and Liatongus may be relatively old dispersal events. Dispersal of an ancestral oniticelline from Africa to Madagascar during the early to middle Cenozoic may be the most likely hypothesis for the presence of the helictopleurines (see Yoder and Nowak 2006) and the clade origin has been dated to between 34 and 21 Mya (Wirta et al. 2008). In contrast and based on the low diversity of the Madagascar onthophagines, their presence there must be due to one or more relatively recent dispersal events, not including the modern accidental introductions.

The New World oniticellines belonging to Attavicinus and Liatongus may each represent separate invasions by ancestral species from Asia to North America via Beringia. The land connection between these continents has been present from the mid-Cretaceous up through the late Pliocene 3.5 Mya and several more times during the Pleistocene (Sanmartín et al. 2001). Cambefort (1991) suggests the colonization of North America by these ancestors is not older than the Pliocene. Based on comparison with probable dispersal events in natricine snakes (Guo et al. 2012), invasions may have occurred even earlier during the late Oligocene or early Miocene ~27 Mya. If dispersal via tropical forest was required, the most likely dispersal event occurred during the Eocene (~50-35 Mya), when the northern hemisphere was much warmer and humid than it is today and a continguous boreotropical forest linked the two continents. At the end of the Eocene, a mixed deciduous hardwood forest changed to conifers and the rise of the Rocky Mountains and Sierra Madre Occidental created a barrier between the western and eastern Nearctic. It is not yet clear if the restricted distribution of these taxa in western North America is due to a mountain high elevation barrier or some other cause.

The Mexican Attavicinus monstrosus generally appears as sister to Paroniticellus, a monotypic genus known from middle Asia and Turkey. This close relationship is difficult to explain compared to the latter’s alternate position in the K = 1 or 2 extreme weighted trees, where it is deep within a paraphyletic Oniticellina. But links between the eastern/western Palearctic with western Nearctic are certainly known (Sanmartín et al. 2001) and therefore a sister relationship is plausible. This lineage is sister to all other Oniticellina and Drepanocerina included in the study, evidence for its antiquity. But no doubt further support is needed for broad acceptance.

The two Nearctic Liatongus species, L. californicus (Oregon and northern California) and L. rhinocerulus (northern and central Mexico) most likely share a most recent common ancestor based on very similar external morphologies and close but disjunct distributions. The one representative included in this study, L. californicus, nearly always appears as sister to a larger clade that has as its basal lineage an Old World species of Liatongus. Zunino (1982), states that these two North American species of Liatongus belong to the Southeast Asian L. phanaeoides group. Thus, the position of the two included species of this genus in this study (one New World and one Old World) as sister taxa in the K = 3 weighted tree may indicate an accurate evolutionary relationship.

The Jamaican endemic Anoplodrepanus and an African representative of Euoniticellus appear as sister taxa in all topologies discovered. It is a relatively derived lineage appearing at the apex of the Oniticellini topology with its sister clade of Cyptochirus + Afrodrepanus. Euoniticellus has an African origin with Palaearctic and Asian species and a single New World member, E. cubiensis (Cambefort, 1996). A link between the faunas of Africa and the Caribbean has been reported elsewhere for many groups of insects and studies give both vicariance and dispersal hypotheses (e.g., Brown 1978, Flint 1977, Liebherr 1988). Both Anoplodrepanus and E. cubiensis may represent two invasions of the New World; one by perhaps by a Euoniticellus ancestral species that reached Jamaica and evolved into what we now recognize as Anoplodrepanus. A second invasion of a Euoniticellus ancestral species is supported by the position of E. cubiensis in a phylogeny of the genus done by Cambefort (1996). From a study on lizards (Gamble et al. 2011), it is possible that these two dung beetle dispersal events happened sometime between 6.0 to 21.9 Mya (but see Wiens et al. 2009 and references within for some contrasting evidence on colonization age and diversification in frogs).

Few biogeographic studies have been completed on individual genera. But in one on Eodrepanus (Barbero et al. 2009b), the oldest lineages are west and central or perhaps east African and a major split between the African and Oriental clades is seen using a variety of analyses (Ochiai similarity matrix, parsimony analysis of endemicity, and dispersal vicariance analysis). Evidence also suggests the effects of an initial African-Palearctic split and later further vicariant events and at least one more dispersal event in Asia.

The Onthophagini dispersal pattern is similar to that seen in the Oniticellini, but continued into Australia and South America, as well as everywhere else where one finds dung beetles. The African, Asian, and Palaearctic regions contain a total of 30, 11, and 3 genera, respectively. But of the total 35 Onthophagini genera, most are endemic to either the African (22 genera) or Asian (4 genera) regions (Davis et al. 2002).

Onthophagus alone is the most widespread as well as the most speciose genus, comprising at least 45% of the species of Scarabaeinae. It is also the only onthophagine genus that has spread outside of the Ethiopian, Oriental, or Palaearctic regions. Based on Emlen et al. (2005), this group has its oldest lineages in Africa and its youngest lineages in Australia and the New World. Dispersal to the New World may have occurred earlier than in the Oniticellini before the rise of the Western Cordillera by the late Eocene, although extensive erosion largely eliminated this barrier by 30 Mya, allowing dispersal between the western and eastern Neactic (Sanmartín et al. 2001). In the New World, the Oniticellini include six species, while the Onthophagini have 139 species (Davis et al. 2002, Delgado et al. 2006). Moreover, Onthophagus spread into South America (95 species) and occur at least as far south as 40th parallel (Cambefort 1991), whereas the oniticellines are completely absent from this region. The high diversity of the onthophagines (compared to the oniticellines) may be due to earlier dispersal and longer time for speciation or perhaps other factors not yet clear.