VOL. 43, NO. I March, 1968 THE QUARTERLY REVIEW of BIOLOGY

LIFE CYCLE ORIGINS, SPECIATION, AND RELATED PHENOMENA IN CRICKETS

BY RICHARD D. ALEXANDER Museum of Zoology and Departmentof Zoology The Universityof Michigan,Ann Arbor

ABSTRACT Seven general kinds of life cycles are known among crickets; they differ chieff,y in overwintering (diapause) stage and number of generations per season, or diapauses per generation. Some species with broad north-south ranges vary in these respects, spanning wholly or in part certain of the gaps between cycles and suggesting how some of the differences originated. Species with a particular cycle have predictable responses to photoperiod and temperature regimes that affect behavior, development time, wing length, bod)• size, and other characteristics. Some polymorphic tendencies also correlate with habitat permanence, and some are influenced by population density. Genera and subfamilies with several kinds of life cycles usually have proportionately more species in temperate regions than those with but one or two cycles, although numbers of species in all widely distributed groups diminish toward the higher lati­ tudes. The tendency of various field cricket species to become double-cycled at certain latitudes appears to have resulted in speciation without geographic isolation in at least one case. Intermediate steps in this allochronic speciation process are illustrated by North American and Japanese species; the possibility that this process has also occurred in other kinds of temperate insects is discussed.

INTRODUCTION the Gryllidae at least to the Jurassic Period (Zeuner, 1939), and many of the larger sub­ RICKETS are insects of the Family families and genera have spread across two Gryllidae in the Order Orthoptera, or more continents. and are identifiable by their jump­ Crickets occupy today all of the major ing hind legs, three tarsal segments, terrestrial habitats between approximately lati­ C and long tactual cerci bearing tudes 55° N and 55° S, or, roughly, the region clumps of knobbed hairs. They range in body between the boundaries of the northern and length from l or 2 mm to over 50 mm. The southern coniferous forests, or taiga (Fig. l). family comprises some 2455 known species, In the course of achieving this broad distribu­ presently arranged in 14 subfamilies and about tion, they have evolved at least seven kinds of 333 genera (Table I). Fossil evidence dates life cycles, most of them more than once in 1 2 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43 various parts of the world. These different .!,"Cl .C:"t:1 ...... -i:: "'01 bO adjustments to climatic variation involve far 0 01 -~; fl d d "'-bl) "[t more than the simple synchronizing of resis­ ...~ ::Id l~ V -a-01 g ..o.~ ~"'..0 8 tant life stages with harsh seasons. Within vO 01 .C: d v~ v· v"t:I ~ v"t:I d O 01 e!'S; complementary investigations in Europe, 01 0 - ~~ s..8 ~s] - ., bl) s ...... Japan, Australia, and North America.

(J This paper summarizes the available infor­ gd ,..;"0-i mation and compares variously related species z -~o..-~ 01 se .., 9 .8 ""; d.., in an attempt to reconstruct some of the evo­ --~ 'ii ;:;s 0.. .c: l::., -~ 8 ~ .., 0:, ... ·;:: instars, or molt ~ -~ .., t:: .c:., .c:!'.i ,; < ..,z ~< ::I ,,, ... ~~ u stages, and the juveniles resemble the adults, r"' l ... ::I animal food of a wide -s" ,,, variety; some tree crickets are believed to sub­ C ,,; el 0 0 - 0 Acheta domesticus (Lin­ .., !!l naeus), Gryllodes sigillatus (Walker) (species ~ !!l < ,

"Trigonidiinae Sword-Tailed Crickets 28 380 trees, bushes, herbs, World-wide small or tiny, delicate, usually pale r,~ leaf litter brown in color t'l .. 0 Cachoplistinae None I 4 soil surface? Ethiopian, Oriental, tiny, soft-bodied :,;;, Australian .....

"Mogoplistinae Short-Winged Crickets 14 200 bushes, herbs, ground World-wide small or tiny, short forewings, en- litter larged pronotum, body with scales ~ ::i.. "Myrmecophilinae Ant-Loving Crickets 5 45 ant nests World-wide tiny, wingless, hump-backed, no audi- § tory organs, short antennae, reduced eyes.

C)~ "Nemobiinae Ground Crickets 16 200 soil surface World-wide small, cf with tibial or tegminal gland s;: "Gryllinae Field and House Crickets 70 550 soil surface and World-wide large size, yellow to black, ~ with ::j shallow burrows long ovipositor 0 ~ "Brachytrupinae Short-Tailed Crickets 12 42 soil surface and deep World-wide resemble Gryllinae except with short burrows (except Europe) or no external ovipositor ~ Q •Gryllotalpinae Mole Crickets 6 65 subterranean World-wide mostly large, heavy bodied; shovelling ..... forelegs, pointed head, cylindrical C) body, dense body setae ~ Totals 14 333 2455 ~

• Subfamilies discussed in this paper.

<:.>o 4 THE QUARTERLY RETIIEW OF BIOLOGY [VOLUME 43

0 0 !:e !Q

0 ~

0 (X)

0

0 <:t

0 (X)

0 ~

0 !£

0 (t) from Asia and Africa, and I have studied vir­ sis Luggar, in Austin, Texas: "During April tually all North American species.] This kind and May the testes, in active spermatogene­ of cycle is rarer than might be supposed. Many sis... " and "At this same period of the tropical and sub-tropical crickets are evidently year the abdomen of the female Myrmecophila cyclic in association with dry and wet seasons. is found to contain a few very large elliptical Wheeler (1900) even implies seasonality in the white eggs. . ." ant nest inhabitant, Myrmecophila nebrascen- 2. One generation each year with egg dia- MARCH 1968] LIFE CYCLE ORJGlNS AND SPECIATION IN CRICKETS 5

Jan Feb

vernalis

fu/toni

veletis

pennsylvanicus

firmus

rubens

assimilis

FIG. 2. DIAGRAMS OF THE LIFE CYCLES OF THE SEVEN FIELD CRICKETS (Gryllus spp.) IN EASTERN NORTH AMERICA, ILLUSTRATING FOUR OF THE SEVEN GENERAL KINDS OF LIFE CYCLES KNOWN IN CRICKETS

Gryflotalpa grylIota/pa

Nemobius sylvestris

FIG. 3. DIAGRAMS OF THE Two-YEAR LIFE CYCLES REPORTED FOR Gryllotalpa gryllotalpa IN SPAIN (Morales, 1940) AND Nemobius sylvestris IN ENGLAND (Gabbutt, 1959) G THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43 pause during winters or dry seasons. Examples life cycle in tropical India for the univoltine are Gryllus pennsylvanicus (Burmeister), Mio­ burrowing cricket Brachytrupes achatinus Stoll, gryllus verticalis (Serville), Valerifictorus mi­ which may be related to this one. Most individ­ cado (Saussure) (introduced) in North America; uals mature in June, but eggs are not laid un­ all North American Nemobiinae north of til September or October; thus juvenile growth about the 30th parallel; most temperate, vege­ is retarded during winter, and the summer dry tation-inhabiting Gryllidae (involving three sub­ season is spent in a prolonged adult stage. families: Oecanthinae, Trigonidiinae, and Ene­ (6) Two generations each year with juvenile opterinae); Teleogryllus commodus (Walker) diapause during winters. Examples are Gryl­ in Australia (Hogan, 1960), T. emma (Ohmachi lus rubens Scudder and G. integer Scudder, and Matsumura) and T. yezoemma (Ohmachi sibling species in North America, Teleogryllus and Matsumura) in Japan (Masaki, 1966). taiwanemma (Ohmachi and Matsumura) and (3) One generation each year with juvenile Scapsipedus parvus Chopard in Japan (Ma­ diapause during winter. The diapause usually saki, I 966). occurs two molts from adulthood. Examples (7) Two generations each year with egg dia­ are Gryllus veletis (Alexander and Bigelow), pause during winters. Examples are Pterone­ G. fultoni (Alexander), G. vernalis Blatchley, mobius fascipes (Walker), P. taprobanensis Anurogryllus muticus (De Geer), Nemobius am­ Walker, and P. ohmachii Shiraki in southern bitiosus Scudder, Falcicula hebardi Rehn, Scap­ Japan (Masaki, 1960, 1961), Oecanthus argen­ teriscus vicinus Scudder (Hayslip, 1943), and tinus Saussure and 0. celerinictus Walker in S. acletus Rehn and Hebard (Hayslip, 1943), in eastern United States (Walker, 1963). North America; Gryllus campestris Linnaeus in Europe (Sellier, 1954); Nemobius yezoensis [The genera discussed above are grouped into six or seven subfamilies as follows: Gryllinae Shiraki in Japan (Masaki and Oyama, 1963) (house and field crickets) = Acheta, Gryllodes, Melanogryllus desertus (Pallas) in southern Gryllus, Melano,rryllus, Miogryllus, Teleogryllus, Russia (Megalov, 1924), Brachytrupes mega­ Valerifictorus (and possibly Anurogryllus and cephalus Lefevre in Tunisia (Valdeyron-Fa­ Brachytrupes, although some evidence indicates bre, 1955), and B. portentosus Lichtenstein in that these genera belong in the subfamily Brachy­ Formosa (Sonan, 1931). In North America, trupinae instead); Nemobiinae (ground crickets) juvenile diapause is unknown in other Ensifera, = Nemobius, Pteronemobius; Trigonidiinae and adult diapause occurs only in two species of (sword-bearing crickets)= Falcicula; Gryllotal­ Tettigoniidae (Copiphorinae), the cone-headed pinae (mole crickets)= Gryllotalpa, Neocurtilla; grasshoppers, N eoconocephalus triops (Lin­ Pentacentrinae (long-legged crickets)= Tohila; Myrmecophilinae (ant-loving crickets)= Myrmeco­ naeus) and Pyrgocorypha uncinata Stal. Al­ phila.] though both species molt to adulthood in the fall, neither is actually sexually mature until The species mentioned, together with those spring, at least in the northern parts of their for which variable life cycles are described be­ ranges (ca. 37° and 39° north latitude, respec­ low, include all crickets for which this author tively). In the far south, some individuals of has been able to find life cycle descriptions. N. triops mature and sing in fall and evidently overwinter as eggs (Walker, 1964). VARIABILITY IN LIFE CYCLES (4) One generation each two years with dia­ pause in both egg and juvenile stage during All of the cricket species with broad north­ winters. An example is Nemobius sylvestris south ranges in the eastern portion of North (Bose.) in England (Richards, 1952; Gabbutt, America exhibit some kind of geographic vari­ 1959). ation in life cycles. Some of these intraspecific (5) One generation each two years with dia­ variations, all of the known cases of which are pause in both juvenile and adult stage during listed below, span wholly or in part the gaps winter. Examples are Gryllotalpa gryllotalpa between cycles and suggest the origins of some Linnaeus in Spain (Morales, 1940) and Neo­ of the differences. curtilla hexadactyla (Perty) in North America (1) Nemobius carolinus Scudder occurs from (Fulton, 1951). Ghosh ( 1912) reports an unusual Canada to southern Mexico and to Key West MARCH 1968] LIFE CYCLE ORIGINS AND SPECIA1'ION IN CRICKETS 7

0. I imits ~porly known ...... ' ··-...... r ...... rt' .. , .. -·

1 \ ...... _____ .,_ .. 1 ,',.

;L.. Gryl/us .. ::\ - -....--·- -.., __ ... ,._~: .. \..,_-.- ..-·- ..... \ 0~ \ ·,• ··· veletis --·-..,_ - ..-·-·- ...... - ..-

adults only

fall adults only

,',> ' ", I :, ".. "., -,"" (or all year in south)

FIG. 4. GEOGRAPHIC DISTRIBUTION OF THREE CLOSELY RELATED FIELD CRICKETS, Gryllus veletis, G. penn­ sylvanicus, and G. firmus Hatched areas represent predicted general distributions. Symbols outside hatched areas represent known peripheral isolates; other symbols are individual collecting records, plotted one to a county. For veletis and pennsylvanicus, only the southern limital records, compiled through the author's field work, and northern limital records from the literature are shown. These two species are the most continuously distributed field crickets within the hatched area, as well as in Washington, Oregon, the Rocky Mountains, and southwestern Canada. Alabama and North Carolina records outside hatched areas are firmus.

in Florida. Evidently, its egg diapause is sim­ range; both have one restricted adult season ply missing in the southern regions where in the North; and both have been taken as adults are present continually. Miogryllus ver­ adults during all months in the South. [UMMZ: ticalis (Serville), extending from central Illi­ records of field observations by R. D. Alexan­ nois and New Jersey to Key West, Florida, der are preserved in the Insect Division of seems to change similarly. Neither species has The University of Michigan Museum of Zool­ any large breaks in distribution across its ogy; specimens, collected by R. D. Alexander 8 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43 and others, are housed in UMMZ collection]. or are derived each year from the juvenile­ (2) Nemobius fasciatus (De Geer), with a overwintering population is unknown. If, north-south distribution almost as extensive however, this species has an egg stage that is in eastern North America as that of N. caro­ to any extent diapause-susceptible or winter­ linus, is believed to produce one generation in hardy, it is clear that the opportunity is being northern states and two in southeastern states. provided repeatedly for the establishment of a Partial third generations may occur at some fall-adult, egg-diapausing population; such a latitudes (Fulton, 1951). Restriction to egg­ derived population evidently could expand overwintering is retained in all but the northward on its own from the southern lati­ southern limits of its distribution where it, tude necessary for its initial separation from too, evidently breeds continually, since adults the parent population. are present at all times of the year (UMMZ). (5) Anurogryllus muticus also extends south­ (3) East of the Great Plains, Gryllus veletis ward from southern Illinois (Hebard, 1934) and G. pennsylvanicus extend southward to and New Jersey (Blatchley, 1920) to at least New Jersey and Maryland along the Atlantic Homestead, Florida (UMMZ). All populations Coast, and to about 34° N in northern Georgia of Anurogryllus in South and Central America, (Fig. 4). In the West, both species extend Mexico, and the West Indies currently are con­ from Canada to central Mexico, probably sidered conspecific with those in North America. with breaks in distribution, particularly in No intensive comparisons have been made, but the northern parts of the Sierra Madre Oriental. the only noticeable difference in preserved ma­ Crickets thus far inseparable from these species terial is a prevalence of macropterousness in in both morphology and song also occur the southern populations and virtual absence around Los Angeles and San Diego, California; of macropterousness in those found in North winter in this region is not comparable to America. The extent of deviations in life cycle that occurring in northern Georgia or the in the southern regions, even in southern western mountains, or even in the central part Florida, is not known. Some periodicity in the of Mexico as far south as San Luis Potosi. appearance of adults occurs there with the The presence of many stages in southern Cali­ largest populations in spring as in more north­ fornia in June, 1963 (UMMZ), indicates con­ ern localities. tinual breeding, but the relationship of the (6) Neocurtilla hexadactyla has an unexpect­ non-wintering populations there to crickets edly large range for a species restricted to thus far indistinguishable from them in the marshes and stream banks. Fulton (1951) has Rockies, and eastward and southward from the indicated that its life cycle at Raleigh, North Rockies, cannot be deciphered without more Carolina, is that depicted in Fig. 3 for G. gryl­ distributional information and further study of lotalpa. Specimens that so far are morphologi­ living material. [The broad distributions of these cally inseparable occur through Mexico and species in some ways resemble the distribution of Central America to Argentina and southern some other species such as the snowy tree cricket, Brazil, and to the southern tip of Florida. Oecanthus fultoni Walker (Walker, 1962), and [This is the basis of resumption of use of the the leopard frog, Rana pipiens Schreiber name originally proposed for a Brazilian speci­ (Moore, 1944).] men by Perty (1832) rather than borealis used (4) Gryllus fultoni extends southward in by Burmeister (1838) (see Alexander, 1964). At Florida, with an increasingly spotty distribu­ this point, it seems important to emphasize the tion, to at least Key Largo (UMMZ; T. J. similarity of specimens from Canada to Argen­ Walker, pers. commun.). North of Florida, tina in this unique case rather than the pos­ this is strictly a juvenile-overwintering, spring­ sibility that more than a single species is maturing species, but in northern Florida represented.] The two Brazilian and two U.S. (Gainesville area) occasional individuals ma­ specimens examined by the author even have ture in fall (Fig. 5). Whether the few individ­ identical numbers of file teeth on the male uals heard chirping there every autumn (T. tegmina, usually an excellent character for J. Walker, pers. commun.) represent a tiny, separating species. It is intriguing that a simi­ self-perpetuating, egg-overwintering population lar two-diapause life cycle has been postulated MARCH 19681 LIFE CYCLE ORTr,TNS AND SPECTATTO,\T TN CRTCKETS

. , . :·"-!,. ;- -.:.- . -~ \❖ ---- ...·

Gryllus fultoni

spring adults only

------a few fall adults t1 ------✓ breeding cycles traced '1/ ·, ~not ' -~-~ . ~~ ~ '\ \·; . ~°' . . f 'i9 '-\~ FIG. 5. GEOGRAPHIC DISTRlllUTION OF Gryllus fulloni Hatched area represents predicted general distribution. Symbols outside hatched area are known peripheral isolates; other symbols are collecting records plotted one to a county.

for another distantly related mole cricket, Gryl­ ica more than .5° latitude farther than the next lotalpa gryllotalpa, in Spain (Morales, 1940). juvenile-overwintering species. All of the sub. Only burrowing cricket species are known to social crickets in North America -Anurogryllus overwinter in other than the egg stage in re­ and the various mole crickets - are either trop­ gions with severe winters. G. veletis and N. ical in origin or evolved their life cycle tenden­ hexadactyla in North America, and G. carn­ cies in south temperate climates and retained pestris, N. sylvestris, and G. gryllotalpa in them as they moved across the tropics to north Europe, extend far north of any other juvenile­ temperate regions. Parental species are not overwintering cricket species - in North Amer- likely to evolve egg diapause; likewise, among JO THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43

Lemperate crickets only juvenile-diapausing or May, 1964, I could hear fewer than six males adult-diapausing species could become sub­ chirping at night on the entire peninsula, social. whereas in August, 1954, 1956, and 1964, thou­ (7) Gryllus firmus is largely made up of a sands of males were chirping, and hundreds of flourishing coastal population extending from adults could be collected along short stretches of Key West, Florida, north around the Gulf Coast the highway. In the vicinity of Anniston, Ala­ to at least Brownsville, Texas, and north along bama, there is a small population of fall adults the Atlantic Coast to the Delaware Peninsula restricted to a field of one or two acres; no (Fig. 4). Southern limits in Mexico are not early summer adults have been found. South­ known, but no males were heard chirping at ward from Anniston, the first small populations night during searches by Alexander either along of spring adults occur at Marianna, Florida. a few miles of the beach and the highway In coastal regions from Carolina Beach south­ behind the beach near Nautla, Vera Cruz (7 ward through Georgia, and on Fripp, Lady, June 1965) or in Tampico, Tamaulipas (22 and Tybee Islands along the coast, relatively October 1966). A series of inland northern small spring adult populations and large fall populations of G. firmus occurs in the belt of adult populations are present. Inland in the sand hills that extends southward from North sand hills of North Carolina, near Spout Springs Carolina into northern Georgia and Alabama in Wake County, the sparse fall population dis­ (Fulton, 1952; Alexander, 1957). This inland covered by Fulton (1952) and confirmed by population is completely isolated from the collections and tape recordings made by Alex­ coastal population in North Carolina, but in ander (1957) has evidently become extinct; the Alabama the inland and coastal populations specific locality has been converted into an are connected by a series of small, restricted extensive pine plantation and I was unable populations. Geographic connections with G. to hear or locate any adults there August 14 and pennsylvanicus, the most similar relative of 15, 1964. No spring adults have ever been G. firmus, occur along the North Atlantic Coast found there. Egg-overwintering fall adults, in Virginia and Maryland and in northern therefore, extend farther north than juvenile­ Alabama. In the first of these two locations overwintering spring adults and are more abun­ there now is evidently overlapping and the pos­ dant where both occur. sibility of interbreeding; the second case in­ (8) 0. quadripunctatus Beutenmuller and 0. volves a series of separated local populations. niveus (De Geer) in eastern North America are File tooth counts and wingstroke rates during univoltine egg-overwinterers in the north but stridulation show some intergradation in Ala­ bivoltine or trivoltine in the south (Walker, bama, indicating a strong probability of past 1962, 1963). interbreeding. During the past ten years I have been accu­ Several cases of life history variations addi­ mulating detailed information concerning the tional to those mentioned above are known in distribution of G. firmus, and the life history crickets from other continents. and morphological and song variations in dif­ ferent parts of its range. Adults and juveniles (I) Gryllus campestris, a univoltine, juvenile­ have been collected or sought in nearly all diapausing species in Europe, is believed to parts of the known range in both early and produce fall adults rarely in southern parts of late summer, and in Florida in mid-winter. its range (Lucien Chop;ird, Franz Huber, pers. Only in southern Florida, below the region of commun.). Lake Okeechobee, is there evidence that all (2) Teleogryllus commodus is a univoltine, stages are present in fairly large numbers at all egg-diapausing species in temperate regions in seasons. In central and northern Florida, some Australia, a non-diapausing species in tropical adults are present at most or all times of the regions (Hogan, 1965). year, but there are definite peaks in early sum­ (3) Scapsipedus aspersus (Walker) in Japan mer and late summer. At Carolina Beach, is a univoltine, egg-diapausing species in the North Carolina, fall adults are abundant, but north, but in certain southern prefectures fac­ the early summer population is tiny: in late ing the Pacific Ocean, an off-season, spring- MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 11 adult, juvenile-overwintering population occurs With regard to variability in life histories, with the fall-adult, egg-overwintering population crickets can be divided into two categories, those (Masaki, 1961). that live chiefly in vegetation and those that (4) Nemobius sylvestris is believed by some live chiefly on or beneath the soil surface. Vege­ European entomologists to have a univoltine tation-inhabiting crickets do not overwinter in life cycle with egg diapause on the European other than the egg stage except in the tropics continent, contrasting with the two-year cycle and in the mildest mesothermal regions. In that Gabbutt (1959) believes to occur in the mesothermal and microthermal regions, egg­ British Isles (See Richards, 1952). overwinterers are rarely multivoltine. Other less spectacular variations have been Four of the seven known life cycles of crickets, indicated in various species by the investiga­ then, are found only in surface and subter­ tions of Bigelow (1958, 1960, 1962) and Raksh­ ranean species in mesothermal and microthermal pal (1962) in North America, Bigelow and regions, and a fifth occurs only rarely outside Cochaux (1962) in Australia, Sellier (1954) and these groups. Two subfamilies, Gryllinae and Fuzeau-Braesch (1963) in Europe, and Masaki Nemobiinae, have evolved six of the seven (1960, 1961, 1963, 1965, 1966) in Japan. Many 0£ known cycles (only the peculiar two-year cycle these, particularly in Japan and Australia, in­ of sub-social Gryllotalpinae is lacking). volve carefully studied polygenic geographic Only one North American species outside variations in intensity and duration of diapause the Nemobiinae, Gryllinae, Brachytrupinae, and and are discussed later in that context. Gryllotalpinae has evolved a juvenile diapause in the mesothermal region. This is Falcicula

DISTRIBUTION OF LIFE CYCLES hebardi Rehn (Trigonidiinae), a grass-clump and leaf-litter species. Two subfamilies char­ The geographic region occupied by crickets acteristic of leaf-litter and surface trash, Mogo­ is approximately bounded on both north and plistinae and Phalangopsinae, occur rarely out­ south by taiga, the coniferous forests of the side the tropics in North America, the former Subarctic and (much less extensively) upland reaching into Ohio (one species, T. J. Walker, South America, as delimited by Thornthwaite pers. commun.), Texas, Arizona, California, and (1933) and others. Within the regions bounded the other into northern Mexico. by taiga, Thornthwaite distinguishes five gen­ Except in the most arid regions, a given eral climatic belts: tropical, north and south locality in the tropics generally has more cricket mesothermal, and north and south microthermal species than a locality of similar size in the (Fig. 1). His classification is useful in dealing mesothermal region; localities in mesothermal with the distribution of crickets and their regions in turn have more species than localities various life cycles. in microthermal regions. South from Ontario, Of the seven known life cycles of crickets, various regions have approximately the follow­ only one, continuous breeding, is prevalent ing numbers of cricket species: Ontario, 12; among tropical crickets. In certain regions, Michigan, 22; Ohio, 33; Kentucky, 34; Ten­ dry-season diapause probably occurs, and devel­ nessee, 35; Georgia, 44; Florida, 67; Grand opmental rates probably vary. Except for hints Bahama Island, 18. in the work of Hogan (1960, 1965) on Teleo­ As one might expect, the proportions of gryllus in Australia and Arndt (1931) on Chre­ vegetation-inhabiting and soil-inhabiting (bur­ mon repentinus Rehn (Eneopterinae) in Haiti rowing and surface) crickets in any given local­ we are, however, as yet provided with no clear ity shift as one moves from microthcrmal to evidence in this regard. In Mexico near Ciudad mesothermal and tropical regions. My field Mante:' (Tamaulipas) and Guadalajara (Jalisco), work in nearly all parts of the United States, I have found that with few exceptions the same Mexico north of Taxco (Guerrero), and on species are adult in June and October, but in Grand Bahama Island suggests that microther­ various localities in the arid region north of mal localities have 50 to 65 per cent soil species; Mazatlan (Sinaloa) on the west coast, only three mesothermal localities, 35 to 50 per cent; and to five cricket species were adult in late June tropical localities, 30 to 45 per cent. A transect but 25 or more were adult in October. southward from Ontario gives the following 12 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43 percentages: Ontario, 61 per cent; Michigan, ferent subfamilies along with non-diapausing 55 per cent; Ohio, 50 per cent; Kentucky, 50 species. Adult diapause, even though it seems per cent; Tennessee, 50 per cent; Georgia, 45 restricted to almost entirely subterranean mole per cent; Florida, 43 per cent; Grand Bahama crickets, may also have evolved more than once, Island, 40 per cent. The Mogoplistinae with for it occurs in European and American species two species on Grand Bahama Island, 14 in in different genera. Florida, two in Georgia, and one in Tennessee, A simple hypothesis is that continuously Kentucky, and Ohio, are excluded because breeding tropical crickets expanding their their leaf litter and surface trash habitat, ranges into temperate regions would have the coupled with the fact that they spend much number of generations between winters (and time on vegetation, makes them difficult to clas­ consequently the number of eventual diapauses sify. Their evident inability to overwinter out­ per generation) determined by their existing side the tropics (except for two species in east­ developmental rates and the growing season ern North America and two species in western length at the latitude where winter survival North America, all of which overwinter as began to be affected. Some species could get eggs) suggests that they should be classified in two full generations, others but a single one, with vegetation inhabitants. Species apparently while still others might have time to move only introduced by man's activities (Acheta domes­ from one diapause-susceptible (or "winter­ tietts, Gryllodes sigillatus, Valerifictorus micado) hardy") stage to another. The last group would also are not included in the above calculations. require more than a year, and eventually (as Evidently the variety of life cycles available to they were able to penetrate far enough north, soil-inhabiting crickets has increased the propor­ or south, and the necessary genetic alterations tions of non-tropical soil-inhabiting species. One occurred), more than one diapause to mature. obvious way is by enabling more tropical soil Reverse changes, it would seem, could produce species to invade non-tropical regions. Another continually breeding tropical crickets from tem­ perhaps less obvious way is through increased perate species having any kind of life cycle speciation in temperate regions, either because involving diapause. of longer inhabitation by more species, com­ This hypothesis probably accounts for many pared to vegetation-inhabiting species, or of the differences among the life cycles of through speciation directly associated with life temperate species, particularly those existing be­ cycle changes. Evidence for both possibilities tween species that are not closely related and exists. In the first case, numerous soil species seem obviously to be derived recently from have evidently penetrated into the mesothermal tropical stock. It does not explain, however, region more or less recently from the tropics why certain stages are evidently more suscep­ (Anurogryllus muticus, N eocurtilla hexadactyla, tible to diapause than others throughout diverse Scapteriscus acletus Rehn and Hebard, S. vicinus groups of crickets, and it does not approach the Scudder, S. abbreviatus Scudder, and others). question of whether or not life cycle differences In the second case, there are temperate sibling ever arise between species restricted to tem­ species such as Gryllus veletis and G. pennsyl­ perate climates, and if so how. It does not, vanicus, with different life cycles, and some therefore, answer the important question of temperate species are divided into populations how it is that juvenile diapause and egg dia­ with different life cycles and temporally isolated pause occur as variations among the species adults (Gryllus firmus, G. fultoni, Scapsipedus within a single genus - a question made more aspersus, and others). interesting by the case of Gryllus veletis and G. pennsylvanicus, sibling species with this particu­

ORIGINS OF LIFE CYCLE DIFFERENCES lar difference in their life cycles (Alexander and Bigelow, 1960) (Fig. 16). How have differences in the life cycles of Two univoltine life cycles are prominent non-tropical crickets arisen? Evidently both among north temperate crickets: (1) egg-over­ egg and juvenile diapause have evolved inde­ wintering with fall adults and (2) juvenile­ pendently several times, since species with each overwintering with spring adults. Both are type of diapause occur in three or more dif- represented in most or all temperate regions, MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 13 and repeatedly involve congeneric species. The ture and extent of polymorphisms correlated most difficult questions about origins of dif­ with habitat permanence and population den­ ferences in cricket life cycles seem chiefly to sity, and the effect of weather fluctuations on involve these two cycles and the kinds of crickets daily cycles of activity. that have them - namely, field and ground crickets (Gryllinae and Nemobiinae) that live THE NATURE OF DIAPAUSE IN CRICKETS on the surface of the soil and excavate shallow burrows. The other four temperate life cycles William Morton Wheeler (1893) introduced are less common, and evidently are restricted the term diapause to describe a stage in the to species derived directly from tropical species morphogenesis of an egg-overwintering meadow or at least lacking close temperate relatives with grasshopper, Conocephalus brevipennis (Scud­ different life cycles. der) (Tettigoniidae: Conocephalinae). Hen­ Comparing various species of Gryllus, Teleo­ neguy (1904) applied the term to the condi­ gryllus (Masaki, 1961), and Nemobius seems to tion of arrested growth represented by the stage suggest only two possibilities additional to those described by Wheeler, and it has since been already given for the origin of life cycle dif­ used in this broader sense to refer to any ferences. First, tropical species moving north period of arrested development regardless of or south, or otherwise being subjected to in­ life-stage and whether associated with winter creasingly severe winters, might give rise to hibernation or summer or dry season aestivation both of the two prominent kinds of univoltine (Lees, 1955). temperate life cycles, rather than just one, as In crickets, non-diapause eggs usually hatch in winter begins to restrict the number of success­ 15 to 25 days at 25° C and with a high degree ful overwintering stages. Second, in certain of synchrony (Masaki, 1961; Rakshpal, 1962; locations or under certain circumstances, tem­ Bigelow, 1962). Diapause eggs characteristically perate crickets with established life cycles might (1) require a few days to reach the diapause produce off-season or off-year populations that stage of morphogenesis; (2) hatch after delay would remain in isolation (temporal, geogra­ and over a long period of time if not subjected phic, or both) and eventually become new to a cold period; (3) hatch with less delay and species. more synchronously if subjected to a brief cold Both of these changes require that certain period, and (4) hatch without delay (in about 15 stages be more diapause-susceptible or winter­ days at 25° C.) and over a very brief period of hardy than others. The repeated, independent time after a long cold period. Diapause adults evolution of diapause in the egg stage and the have not been studied in crickets. Diapause penultimate juvenile stage in different groups juveniles characteristically reduce their activity of crickets and on different continents already and their metabolic rate during the three im­ makes this a certainty. mediately pre-adult instars, and mature in ways The development of two life cycles from similar to the hatching performances of dia­ one stock, or one overwintering stage, seems pause eggs when subjected to similar conditions. most likely to occur in the colder part of the In non-diapause juveniles the later stages are range of a mesothermal species and in the warmer part of the range of a microthermal similar in length to the earlier stages (Bigelow, species. These would be the places where in the 1960; Rakshpal, 1964; Ohmachi and Masaki, first case winter would be severe enough to 1964; Fuzeau-Braesch, 1965). cycle the population, and in the second case winter would least likely be so severe as to !>IAPAUSE SUSCEPTIBILITY OF DIFFERENT annihilate off-season individuals. LIFE STAGES Before considering in detail the relationship of life cycle differences to speciation, it is appro­ Numerous aspects of potential winter-hardi­ priate -to review the available information on ness may be cryptically involved in the develop• several other phenomena associated with life mental physiology of all crickets, even those cycle variations: the nature and intensity of living in tropical regions, but as yet we have no diapause and its geographic variations, the na- direct knowledge of any such characteristics. 14 THE QUARTERLY REVIEW OF BIOLOGY LVoLUME 43

Some other possible explanations of restrictions ing by late juveniles and females is prominent in over-wintering stages, however, are evident. only in species with burrowing males and may Overwintering in other than the egg stage be restricted to them. Burrowing by the late occurs only in surface and subterranean species. juvenile stages could have been selected as an Three stages in such species - eggs, late ju­ overwintering advantage, and also in the same veniles, and (to some extent) adults - are pro­ context as that of the adult male, if newly tected from climatic extremes by being under­ emerged males with burrows already excavated ground, just as eggs of vegetation-inhabiting have an advantage over males that have to species are somewhat protected by injection into begin excavation after maturing. The burrow plant stems. Eggs are buried by the ovipositing of a stridulating male may also be important females, and the two other stages ensconce in protecting him from predators responsive themselves in burrows. These are the only to his stridulation. This idea is substantiated three stages that overwinter in northern cli­ by the close association of calling males with mates. But the same three stages are largely burrows or crevices and their rapid and direct subterranean in the same kinds of crickets in retreat into their burrows upon disturbance. tropical climates, indicating that the selection In some field crickets, one often finds up to occurred in other contexts and constituted a three adult females in the bottom of a stridu­ pre-adaptation as far as winter survival is con­ lating male's burrow, and West and Alexander cerned, rather than an a posteriori response to (1963) suggest that females of the burrowing, winter conditions based on some kind of physio­ sub-social species, Anurogryllus ·muticus, may logical cold-hardiness restricted to certain stages. often take over a male's burrow for the brood In other words, any physiological specializations chamber. It would seem that burrowing by presently making these stages unusually winter­ adult males has to some extent pre-adapted hardy seem, on the basis of available evidence, field crickets both for increasingly subterranean appropriately viewed as adjuncts to behavioral life involving both sexes - as now exemplified characteristics originally responsible for the by Gryllotalpinae and Brachytrupinae - and restriction of selection for physiological winter­ for overwintering in the late juvenile stages. hardiness to these particular stages. The only The early and intermediate juvenile stages are alternative is the exceedingly remote possibility in many species the only stages that are not that all burrowing by late juveniles evolved in more or less continually associated with a temperate regions in juvenile-overwintering burrow. species. In northern climates, only mole crickets over­ Nearly all crickets "bury" their eggs, either winter as adults. They seem to have been in the soil or in plant stems; the known ex­ assisted toward this condition by two specializa­ ceptions are a few sub-social species that drop tions: their elaborate subterranean life, includ­ their eggs in their burrows without injecting ing feeding on such things as earthworms and them into the soil (West and Alexander, 1963). roots, and the lengthened life of adult females, Oviposition into the soil is ancient in the a feature that in arthropods is commonly asso­ Orthoptera, apparently extending to Paleozoic ciated with parental attention to eggs and cockroaches, some of which evidently had long, young. tubular ovipositors (Laurentiaux, 1951). Long ovipositors may never have been lost in some GEOGRAPHIC VARIATION IN DIAPAUSE of the phyletic lines leading to the various groups of modern surface-dwelling crickets. Some of the cases of life history variation Adult male field crickets characteristically described above undoubtedly involve geographic burrow more extensively than females, evidently variation of a genetic nature. As examples, a in association with attracting the more "roving" genetic basis for variations in both egg-hatching females by long-continued, stationary stridula­ time and post-hatching development time has tion. Mate attraction by stridulation is ancient been demonstrated by Masaki (1961, 1965) in a in the Suborder Ensifera, pre-dating the family Japane_se egg-over-wintering Teleogryllus spe­ Gryllidae, and probably beginning sometime in cies; and Hogan (1965) has shown experi­ the Triassic Period (Alexander, 1962). Burrow- mentally that the tropical, non-diapausing MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 15 populations of T. commodus in Australia differ omission that often renders the results useless. genetically from the temperate, egg-diapausing Two kinds of geographic trends in diapause populations. But photoperiod regimes have also variation have been demonstrated in crickets. been shown to be instrumental in initiating Among species restricted to cold temperate or by-passing diapause or altering develop­ regions, diapause intensity and duration in­ mental time in juvenile-overwintering crickets crease in the direction of milder winters (Bige­ (Fuzeau-Braesch, 1963; Masaki and Ohmachi, low, 1960; Masaki, 1961, 1965). Species which 1967; Fig. 6). Accordingly some geographic extend from the subtropics into warm tem­ variation in diapause expression could exist - perate regions show a reversed trend, with in fact, must exist in any cricket widely dis­ populations in the subtropical regions lacking tributed in a north-south direction - indepen­ diapause and those at higher latitudes having dent of genetic variation and owing solely to progressively more intense diapauses in the di­ the larger shift in day lengths during the season rection of more severe winters (Hogan, 1965). at higher latitudes. Among other things, such Clinal variation and intermediacy in hybrids information places stringent requirements upon indicates polygenic control of such variation in experimenters that they might not have con­ the cases studied (Masaki, 1965). Hogan (1965, sidered. Many of the numerous rearing tests on 1966) and Bigelow and Cochaux (1962) found crickets do not report photoperiod data, an that the proportion of diapause eggs of T. com-

85 o = long day; macropterous d' • = short day; micropterous X = short day; macropterous 80

CJ) • C • :2 X 0 75 -0 ::r:... • • • Cl) 70 • 0 • • • 0 • -0 ••• ->, 0 65 .i: 0 -:, •• -0 :i; 60 0 • 00 • •••o •• 0 0 II) • .. x 00 00 - • • 00 >, eoo 0 0 55 • 0 0 000 0 • 0 0 0 0 0 0 0 0 0 50 0 0 oo 45 15 20 25 15 20 25 Body Length in mm.

FIG. 6. EFFECT OF PHOTOPERIOD ON DEVELOPMENT TIME, WING LENGTH, AND BODY LENGTH IN Gryllus integer, A BIVOI.TINE, JUVENILE-DIAPAUSING CRICKET Photoperiod was 12 hours each 24 hours until 18 days after hatching; then 14 hours for long-day, 12 for short-day, with 15 minute increases and decreases, respectively, each five days until the tests ended. Temperature was 82°F in both incubators, with six one-degree deviations downward in the long-day incubator, and three one-degree deviations upward and two downward in the short-day incubator. 16 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43 modus gradually increases toward the south direction of selection may make the slightest in Australia. Hogan (1966) believes that the break in geographic continuity highly signifi­ capacity to diapause may be monogenic in this cant. Fulton (1951) suggests that in the widely species, but that the intensity of diapause may distributed species, Nemobius fasciatus, multiple be polygenic or cytoplasmic. Tendency not to generations also first appear at about these diapause seems to be dominant, and Hogan sug­ latitudes. gests that mass introduction of non-diapausing It would seem from this evidence that some tropical individuals into temperate regions species, particularly those that have occupied might be an appropriate control. He refers to their present ranges relatively recently, would crickets from different regions as different reveal their origins (whether northern or south­ "races," however, and he and Bigelow and ern) by the nature of geographic trends in the Cochaux (1962) found some inter-sterility be­ intensity and duration of diapause. A re-exami­ tween these "races." Hogan (1967) found evi­ nation of species distributions leads one to di­ dence in the laboratory of differential mating vide the crickets of the world into six general that reduced hybridization between diapausing categories: (I) strictly tropical and sub-tropical and non-diapausing populations, and was ap­ species, (2) chiefly microthermal species (with parently eliminated by silencing the males. How more intense diapauses toward the lower lati­ these findings relate to field conditions is still tudes), (3) species extending from tropical or unknown. subtropical regions into warm temperate (meso­ The selective bases for the various geographic thermal) regions (with more intense diapauses trends in the presence or absence, intensity, or toward the higher latitudes), (4) chiefly meso­ duration of diapause are not difficult to imagine. thermal species, (5) species extending broadly At high latitudes in the temperate region, winter across both mesothermal and microthermal re­ temperatures do not rise above the develop­ gions, and (6) species extending from tropical mental threshold for crickets, and the autumn or sub-tropical regions into microthermal re­ temperature-drop and spring temperature-rise gions. In crickets, geographic trends in dia­ probably occur across shorter time periods and pause intensity have not been investigated in with fewer fluctuations than in the milder the last three kinds of species. temperate regions. Winters that are more fickle and are associated with longer summers, yet POLYMORPHISMS, UFE CYCLES, AND HABITAT remain sufficiently severe to be annihilatory to PERMANENCE all but the diapause stage, favor a more positive diapause. On the other hand, gradual penetra­ Hatching time and development time are tion of a subtropical cricket into a region with only two of many characteristics that vary in winters annihilatory to all but one or two ways significant to the understanding of adapta­ stages would necessarily be accompanied by tions by crickets to different climatic conditions. selection producing not only a higher propor­ As illustrated in Table 2 and Figs. 6-1 I, body tion of diapausing individuals but an increas­ si1e, wing length, body form, pigmentation, ingly intense diapause. aggressiveness, and propensity to burrow also These facts yield an interesting situation. vary, not only between species, but within Diapause intensity and duration, at least in species as well. Interspecific differences must species that are not recent immigrants, tend to in most cases reflect genetic variation, one way decrease in both directions from a region of or another, particularly when the species in­ mild winters that seems to fall around 33° N to volved are sympatric and synchronic and when 37° N. Northern Florida and the southern the differences persist in laboratory cultures. tip of the Appalachians are regions where Except for Masaki's findings (1967) concern­ changes in life cycle are associated with changes ing size variations in Teleogryllus emma, how­ in morphology and behavior in numerous Gryl­ ever, the observed intraspecific differences seem Iidae and Tettigoniidae. One is led to believe rarely to be related to genetic variation. In­ that these are somehow "fragile" connections stead, they can be produced by adjusting one between northern and southern components of or more of various natural and artificial stimuli widely ranging species, where changes in the (Figs. 10, II). MARCH 19681 LIFE CYCLE ORIGINS AND SPECIATTON IN CRICKETS 17

TARLE 2 Habitat, life cycle, and proportion of macropterous individuals in 13 field and ground crickets from North America, Europe, and Asia

APPROXIMATE % MACROPTER- NUMBER OF OVERWINTERING GENERATIONS OUSNESSIN SPECIMENS SPECIES HABITAT STAGE PER YEAR FIELD EXAMINED

GryllllS vernalis 1 woods juvenile 1 0 ca. 200 .. fultoni 1 .. " I 0 ca.300 Nemobius sylvestris " " ½ 0 ?(Sellier,1954) .. tinnulus 2 " egg I 0 ca.600 .. yezoensis "in and around juvenile I 0 100 paddy fields" Gryllus campestris grasslands " I I ?(Sellier,1954) Miogryllus verticalis " " I l ca. 1000 Gryllus veletis • " " l 4 314 " pennsylvanicus • .. egg l 4 1000 Nemobius fasciatus " .. 1-2 10 ca.1000 " allardi • " " I 25 ca.1000 Gryllus rubens .. juvenile 2 8 &:75 50 &:150 .. firmus (spring) " " 1 14 107 " firmus (fall) .. egg 1 18 116 Superscripts indicate three pairs of sibling species.

Pigmentation crickets reared in isolation unless the cage is especially rigged to transmit vibrations made Variations in pigmentation extending from by the movements of the other species. almost completely yellow individuals to com­ pletely black ones are geographic in G. firmus and correlate with soil color; similar variations Wing Length can be induced by different amounts of crowd­ ing in G. bimaculatus. Fuzeau-Braesch (1960, Wing length variation is probably the easiest 1961, 1962) and Levita (1962) reported that, in characteristic on which to base a discussion of the second case, tactual, visual, and olfactory polymorphism in crickets because its general stimuli all seem to be implicated, according to significance is clear (Alexander, 1961), and it the following evidence: correlates closely on an interspecific basis with (l) Crickets reared in isolation are darker variations in habitat and life cycle. Further• than crickets reared in groups. (2) Crickets more, with the exception of pigmentation, all reared in isolation in the dark are darker than of the other polymorphisms seem to be related those reared in groups in the dark. (3) Crickets in some fashion, either directly or indirectly, to reared in groups in the dark are darker than wing length variations. those reared in groups in the light. (4) Crickets reared in isolation in a cage with walls made [As used here, the term macropterous means of mirrors are lighter than those reared in that the hind (under) wings extend beyond the forewings or tegmina; some individuals are more isolation in cages with plain walls. (5) Crickets macropterous than others (Figs. 7-9) and fly reared in groups with their antennae and cerci easily, some can fly only a few feet or yards. severed are darker than normal crickets reared Micropterous means the hind wings are com• in groups. (6) Crickets reared in cages with pletely hidden beneath the tegmina; these indi­ other insects (such as grasshoppers) are like viduals cannot fly, and usually do not open their 00

~ ;o ~ ::i:,

~t--< ~ ::i:, t":1 ~ ~ 0 ">1 ...... b:i 0 t--< 0 ~ ~

FIG. 7. MICROPTEROUSNESS AND MACROPTEROUSNESS IN Gryllus veletis Left, a very short-winged adult male, unable to stridulate, found in a laboratory population; center. the usual micropterous male with normal tegmina; right, a macropterous male /note tips of long underwings), produced in the laboratory. Only the last two are known in field populations. < 0 8 s: Pl t; MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 19

Frc;. 8. MrcROPTERor;s;s;r-:ss A:-

hind wings at all, or lift their tegmina except of grasslands that occur over the same area - prior to stridulation. Some variation in tegminal now chiefly lawns, pastures, and old fields, all length is associated with these drastic variations of which are man-made situations that possibly in length of the hind wings, but individuals represent pioneer associations in the pre-agri­ with greatly reduced tegmina are usually re­ cultural landscape in these regions. Certain ferred to as neotenic (Fig. 8); again, two degrees of neoteny are known, insofar as tegminal size is kinds of marshes and prairies in North America concerned. In some crickets, the general body also represent permanent habitats and contain shape, particularly head and pronotal conforma­ wholly micropterous species, such as Nemobius tion, is changed rather strikingly by differences meloclius (Thomas and Alexander, 1957); this in wing length in ways that resemble geneti­ may explain the presence of the micropterous cally-based differences among species (Alexander species, N. yezoensis, in rice paddies in Japan. and Thomas, 1959).] But variations in habitat permanence are not All of the four woods-inhabiting species listed sufficient in themselves to explain the observed in Table 2 are totally micropterous in the field, variations in macropterousness. There is also regardless of overwintering stage and number of a relationship with the kind of life cycle that generations per year; all of the grassland species, is evidently even more fundamental. Thus, on the other hand, produce some macropterous the seven strictly juvenile-overwintering species individuals. In eastern North America, this in Table 2 average less than l per cent mac­ general relationship extends, with exceptions, ropterous, while the four strictly egg-overwin­ to Tettigoniidae and Caelifera· as well. Perhaps terers average nearly 10 per cent macropterous. this relationship could be predicted on the Among the seven grassland species (excluding basis that woodlands are more permanent fiabi­ N. yezoensis), the three univoltine juvenile­ tats, at least in North America, than the kinds overwintering species average only two per cent 20 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43

N. yezoensis is the only ground cricket in Japan that overwinters as a juvenile, and also the only one in which only micropterous individuals are known in the field: " ... both long-winged and short-winged specimens are usually found in the related species." Long photoperiods, either with or without cold treatment, produced rapidly developing individuals, up to 50 per cent of them macropterous in the laboratory. Gryllus veletis and G. pennsylvanicus repre­ sent an apparent exception to the rule that differences in wing length occur between similar juvenile-overwintering and egg-overwintering species, since about 4 per cent are macropterous in both cases (Table 2). It would be interesting to compare northern and southern samples of these species, as well as early and late-maturing individuals. Of seven individuals of veletis collected 27 May 1965, only the last two to mature were macropterous; presumably they were subjected to longer day lengths at com­ parable developmental stages than the early­ maturing individuals. Under continuous light in the laboratory, nearly all individuals of veletis develop directly, without evident dia­ pause delay, and in contrast to crickets reared under ordinary day lengths most (13 of 19 in the only test) are macropterous. Further, there is a concentration of micropterousness among the latest of such individuals to mature. Ju­ FIG. 9. A FULLY MACROPTEROUS MALE OF Gryllus veniles of pennsylvanicus were collected 28 rubens WHIRRING His LONG UNDERWINGS JUST PRIOR TO FLIGHT August 1966, and reared in three groups at 75° F.: (I) six-hour days, (2) normal day lengths, macropterousness in the field, while the three and (3) continuous light. All mat~red at the egg-overwintering species average 13 per cent. same times and all were micropterous. Evi­ Further, in the single bivoltine species, C. dently, day length has its principal effect on rubens, the overwintering juveniles are 8 per younger individuals of this species. Two possi­ cent macropterous, while the directly maturing ble influences on the correlation between wing generation is 75 per cent macropterous. A length and overwintering stage are evident. comparison of 107 pre-July adults and 116 post­ First, juvenile-overwintering populations are July adults of Gryllus firmus, representing the characteristically less dense, and Fuzeau-Braesch two seasonally separated populations of this (1961) and Saeki (1966a) have shown this can species that overwinter as juveniles and eggs, be a factor in some species. More significantly, respectively, gives figures of 15 macropterous long days have been shown to produce long individuals among juvenile-overwinterers and wings in both univoltine and bivoltine species, 21 macropterous individuals among egg-over­ while short days produce short wings (Saeki, winterers, or approximately 14 and 18 per cent 1966b; Masaki, 1966; Masaki and Oyama, 1963; respectively. Thus, even within species, there Fig. 6). may be more macropterous individuals among Wing polymorphism in crickets has been juvenile-overwinterers than am011g egg;0ver­ shown to be related to other phenomena which winterers. Masaki and Oyama (1963) note that affect or correlate with development time, al- MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 21

-- Hereditary-variatians7 (mostly interspecific) PIGMENTATION ---,JBODY FORM I associatedchanges associatedt changes

BODY SIZE affects long-wmged a'ssociatedchanges • - - ['i~-utniiooJ-' ' , individualsfly influenfes / _ - - - affects DEVELOPMENTTIME --- _ I 1 leads to J~~~~~~redu/ 1------L_,___ l lengthenscertain stages dll'ectlyaffects "'"'"-7""'1 I I I I LI prevents I ---- influencesonset of --:·----< '---~ ...... _~ I /---~-\.. r,_.,r..,! ~-~-~, ' Temperature : influences? I

I I changesamount of influences? ' 'I

BRAIN HORMONE (all mutuallyrenforcing) :~~,

AGGRESSIVENESS

Hereditary variations L ( most!~i~!~rs~~f!:~- F1c, 10. A SUMMARY OF THE lNTER·RELATIONSHIPS OF DIFfl•:RI,NT MORPHOLOGICAL, PHYSIOLOGICAI., AND BEHAVIORAL VARIATIONS IN CRICKETS The nine characteristics enclosed in heavy-lined rectangles vary interspecifically, presumably at least in part because of hereditary differences. They are also adjusted intraspecifically in the ways indicated, by differences in crowding, nutrition, temperature, and photoperiod. Sources of evidence are given in the text. Dashed lines indicate probable relationships.

though all of the reported results do not seem wing growth; that with respect to wing growth to agree. Thus, Sellier (1954) found with vari­ the threshold to stimulation from the brain is ous European species that neither adjustments higher in some species than in others; and that of temperature, light, and humidity, nor castra­ the sporadic appearance of macropterous indi­ tion produced macropterous individuals. On viduals in G. campestris can be attributed to the other hand, acceleration of development an abnormal functioning of the brain. Sellier by preventing the onset of diapause in the also reported that Acheta burdigalensis Latreille ninth instar did cause macropterousness. This produces a higher percentage of macropterous was done by massive injections of brain tissue individuals under high, even temperatures than from younger (not yet diapausing) juveniles of under low, fluctuating temperatures. G. campestris, or from juveniles of G. bima­ In a test conducted by the author with G. culatus (which does not have a juvenile dia­ integer, in which all photoperiods were ten pause). Sellier concluded that the brain pro­ hours each 24 hours, at 33° C, 87 females ma­ duces a non-specific "morphogenic" substance tured 35 to 59 days after egg-laying and were (presumably the "brain-hormone" of other au­ 98.8 per cent macropterous; 79 males matured thors) which prevents diapause and induces in 38 to 56 days and were 85 per cent mac- 22 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43

little emigration much emigration and colonization and colonization

increasing hkehhood increasing likelihood of males attracting of moles and females females from long meeting by chance range decreasing likelihood decreasing likelihood of moles attracting of males and females meeting by chance females from long range

less dispersal more dispersal\ ( by flight by flight

more colling by moles t \ less colling by males decreased increased t macropterous­ macropterous- t ness ness development of no territoriality in males fl territoriality in males

~ long periods tofimmobility long periods timmobiltt:, in isolation \ in contact wittother crickets

( tendency to ttherfight or ) tendency to become passive retreat upon touch ) upon touch

( mdi~duals liated,"""'Y individuals c!wded, rarely contacting other crickets out of contact with other crickets

Low-Density Populations High-Density Populations

FIG. 11. SOME EFFECTS OF POPULATION DENSITY ON BEHAVIOR AND MORPHOLOGY OF FIELD CRICKETS AND How THEY ARF.REFLECTED IN THE PERFORMANCE OF THE POPULATION

ropterous. At 25° C and the same photoperiod, 82 to 129 days and were 84.2 per cent mac­ 24 females matured in 9 l to 243 days and were ropterous. Mortality in the three tests (sexes 87.5 per cent macropterous; 21 males matured combined) was 52.8, 76.3, and 72.5 per cents, in IOI to 235 days and were 85.7 per cent respectively. These results indicate that in the macropterous. Seventeen females reared at 25° C laboratory, low even temperatures (I) retard until 59 to 97 days after egg-laying, then development, (2) cause maturation to be less changed to 33° C, matured 78 to 128 days after synchronized, (3) cause higher mortality, and egg-laying and were all macropterous; 19 males (4) increase micropterousness, at least in fe­ reared under the same conditions matured in males. A shorter photoperiod might have exag- MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 23 gerated the last effect. The selective basis for tially the same effect on wing length. All of the lower macropterousness in males may be that necessary experiments to test this hypothes:s most emigrating females are inseminated, reduc­ have not yet been carried out, but the data re­ ing the relative reproductive chances of cmigrat• ported above seem to support the likelihood that ing males. it is true in at least some cases. Particularly Evidently, Gryllus desertus is the only case in pertinent are the findings of Fuzeau-Braesch which genetic variation is believed to account ( 1961) that in two of six species she tested, for an intraspccific variation in proportions of rearing in groups, as contrasted with rearing macroptcrousncss in crickets (Scllicr, 1954), and in isolation, caused significant increases in it seems reasonable, based on the paucity of the proportion of macropterous individuals. evidence of a taxonomic nature, to question Saeki (1966a) obtained similar results with whether the southern populatiops that arc Scapsipedus aspersus. macroptcrous are conspccific with the northern Chauvin (1958) discovered that raising house micropterous populations. Only two other cases crickets (A. domesticus) in pairs or groups ac­ of geographic variation in proportions of mac­ celerates development, an effect erased by re­ roptcrous individuals seem to have been de­ moving the cerci and antennae, suggesting a scribed in insects. Lindroth (1949) believes that tactual stimulus such as Ellis (1959) found in such variations in ground beetles (Carabidae) migratory locusts. Fuzeau-Braesch (1965) ob­ reflect genetic variation and cites laboratory tained similar results with G. campestris. Mc­ tests and the prevalence of long-winged indi­ Farlane (1966) has summarized his results, simi­ viduals at the periphery of the ranges of poly­ lar to the above, obtained with Gryllodes sigil­ morphic species as evidence for this view. He latus (Walker), and has indicated that a does not provide evidence, however, that photo• fundamental and pervasive relationship between period was controlled in the laboratory test, developmental rate and wing length exists and he does not discuss the possibility that through many or all groups of crickets. peripheral populations produce more long­ winged individuals solely because of the trigger­ ing effects of more rapid fluctuations in habitat. Body Size Thomas and Alexander (1962) demonstrate geographic variation in proportions of long• Variations in the body size of crickets have winged individuals of three meadow grasshop• been reported by various authors to be asso­ per species in the Orchelimum concinnum ciated with (I) differences in latitude and alti­ group. Only micropterous individuals arc pro­ tude (Alexander and Bigelow, 1960; Masaki, duced in relict areas, while the flourishing 1966, 1967); and (2) variations in devel­ populations in other regions produce many mac• opment time caused by photoperiod or tempera­ ropterous individuals. No cvidcnc~ is available ture differences (see Fig. 6, and Masaki and to test the question whether this variation is Oyama, 1963). Masaki (1966, 1967) has made owing to one or the other of two apparent an intensive study of altitudinal and latitudinal possibilities: (1) genetic change resulting in loss size variation in two Japanese crickets, Teleo­ of the ability to produce macropterous indi­ gryllus emma and T. yezoemma. He discovered viduals in the relict area; or (2) failure of relict that size differences persist when strains from populations to reach densities necessary to trig­ different regions are reared under identical con­ ger macropterousness. ditions in the laboratory. Short photoperiods Alexander (1961) noted the close correlation caused faster development and smaller adults between wing length, flying ability, and habitat in these egg-diapausing species; faster develop• permanence among different species, wondered ment and larger adults, on the other hand, re­ if behavioral events such as might result from sult from the effect of long days on juvenile­ crowding could affect wing length, and specu­ diapausing species (Fig. 6; Masaki and Oyama, lated that anything which accelerates develop­ 1963). Fuzeau-Braesch (1966) found that dia­ ment, such as highly nutritious diets, high even pausing individuals of G. campestris are larger temperatures, breaking of diap;mse, and tactual whether they mature before or after non-dia­ stimulation from crowding, might have cssen- pausing individuals retarded by low tempera- 24 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43 tures during early juvenile stages. Because eg·g changes ought to be less effective in tropical size seems always the same, Masaki believes that regions); (2) the method of pair formation may in T. emma larger females lay more eggs, and he adjusted, in some fashion compensating for that, omitting other possibilities such as preda­ the lower density of adults in populations; and tion, selection favors larger size except when (3) longer adult life may be advantageous, on maturing time causes eggs to be laid so late the basis that longer-lived adults in a sparse that they fail to reach the diapause state before (adult) population will have a better chance the onset of winter. Likewise, small females to reproduce maximally, and also because suc­ maturing and laying eggs too soon in southern cessful offspring can be produced during a climates run the risk that the eggs will hatch greater part of the year (than in a temperate before winter. There are additional compensat­ region where overwintering stages and develop­ ing factors: high temperatures intensify egg mental habitat are restricted). Pace (1967) diapause, thus reducing likelihood of premature suggests that long-lived adults may generally he hatching in carly-laicl eggs; diapause is more more prevalent in temperate species that over­ intense in southern populations; and the photo­ winter in several stages and live in habitats period effect adjusts maturation times of the continuously suitable for development. juveniles. In at least some cases, longer develop­ ment time and larger size are associated with an exlra molt (Saeki, 1966b). ( ;,yllus rn111p,,stris

Fuzeau-Braesc.:h (19G3) studied the role of Length of Adult Life photoperiod in adjusting the life cycle of this univoltine, juvenile-overwintering species. She Winter may be a cause of high mortality in found that juveniles burrowed more when iso­ northern crickets, but it is also a convenient lated, and more when photoperiods were long synchronizer of adulthood. In species in which · (up to about 20 hours). Eight-hour photoperi­ adults tend their young, or those in which there ods speeded up the molting from the eighth is little competition between adults and juven­ instar into the ninth diapause instar; 16-hour iles (for example, holometabolous insects), long­ days retarded it; and continuous light gave lived adults may by more prevalent than in intermediate results. Short photoperiods during species in which the adults compete with their egg laying also seemed to delay molt of the own offspring. Temperate crickets which do not offspring to the diapause stage. She suggested tend their young evidently all have short-lived that burrow inhabitation effectively shortens adults (Figs. 2, 11). In contrast, adults of tropi­ the cricket's photoperiod, thereby hastening the cal species seem, on the basis of preliminary ob­ onset of diapause. This would mean that servations in the laboratory, to live many times individual crickets which had excavated suitable as long as temperate adults; and most tropical overwintering burrows would enter diapause and sub-tropical species seem to be represented more quickly, thus probably reducing their by at least a few adults, often simultaneously activity, their nutritional requirements, and with juveniles of many ages, at all or nearly all their likelihood of losing contact with the bur­ times of the year. row while locating food. In contrast, molt and A given area habitat unit cannot produce cliapause would be delayed among crickets with­ as dense an adult population of a continuously out suitable burrows. The synchrony of adults breeding species in which adults and offspring that is so evident within cricket populations compete as it can of a seasonally cyclic species in (Fig. 12) thus seems, in spring-maturing species, which they do not. There can never be as many to have a significant beginning during the pre­ adults around in the former case. Three possible vious summer. selective trends in the tropics may be considered As already mentioned, G. campestris produces in this general context: (1) selection for syn­ macropterous individuals extremely rarely, evi­ chronized adulthood may be based on some dently only when diapause is by-passed for one regular seasonal change such as dry and wet reason or another. Whether this capability is seasons (both temperature and photoperiod being maintained by selection or is a vestigial MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 25

300

frl Mostnights and daysbelow 50°F. ~ 8200

Cl) w ...J

~ 100

Cl) a: ~ ~ z:)

heavy frosts

MAY JJNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER FIG. 12. SEASONAL RELATIONSHIPS OF ADULT POPULATIONS OF Gryllus veletis (LEFT) AND G. pennsylvanicus (RIGHT) IN A BRUSHY PASTURE IN SOUTHEASTERN MICHIGAN, AS MEASURED BY COUNTS OF CHIRPING MALES High counts were on wann, sunny days; low counts were on cold, cloudy or rainy days. characteristic is unknown, but the same rare on developmental time in G. integer, a macropterousness occurs in two other univol­ sibling to G. rnbens which has the same bivol­ tine, juvenile-overwintering species, M iogryllus tine life cycle (Fig. 2) and was available when verticalis and Nemobius yezoensis. In all these rubens was not. Long increasing photoperiods cases, females are more likely to be macropterous caused more rapid development and 100 per than males. cent macropterousness, while short decreasing photoperiods retarded development and caused almost 100 per cent micropterousness (Fig. 6). Gryllus rubens and G. integer A serious shortcoming of previously published results was failure to provide adequate tempera­ Alexander (1961) has noted that, based on ture data, particularly since most lights produce collections, proportions of macropterous and heat, and higher temperatures and long days micropterous individuals are almost exactly have similar effects. Temperatures were con­ reversed in the spring and fall adults of the trolled by automatic thermostats throughout bivoltine, juvenile-overwintering cricket, Gryl­ these tests, and adjusted so that the slight over­ lus rubens. It was suggested that development all temperature difference between cages would time might be involved, since one generation tend to favor the crickets kept under short (8 per cent macropterous) is composed of dia­ days. pausing, overwintering individuals, while the The results of these tests are like those of other (75 per cent macropterous) matures with­ Masaki and Oyama (1963) and Masaki (1966) out such delay. with Nemobius yezoensis and Teleogryllus I have since tested the effect of photoperiod yezoemma, univoltine species, in that individuals 26 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43 reared in short days were also smaller than also affect embryonic development in the egg, those reared under long days. either directly or via the mother, but no evi­ G. rubens and G. integer live in lawns, pas­ dence is available regarding crickets. In egg­ tures, roadsides, and old fields, suggesting that diapausing crickets photoperiod thus far seems their pre-civilization habitats were relatively only to time the period of egg-laying by adjust­ unstable. Retention of flight and emigrating ing maturation time (Masaki and Ohmachi, (colonizing) capacity will in the long run be 1967). favored in such nomadic species if habitats deteriorate and disappear. Even though emi­ POLYMORPHISMS, POPULATION DENSITY, grating individuals may not be reproductively AND COLONIZATION favored at the time of emigration or during the generations immediately following it, the fact Some phenotypic variations in insects are remains that if the original habitat eventually density-dependent. Migratory locusts are an disappears, only the descendants of migrants are especially well-known example. Contrary to left. the conclusions of some biologists discussing These species also have to contend with this general topic, I think the evidence sug­ fairly severe winters, at least in the northern gests that density-dependent polymorphisms parts of their ranges. The significance of their tend to increase the reproductive capacity of complex phenotypic variations may be sum­ the individuals possessing them except in cases marized as follows: most late summer juve­ when insufficient time has elapsed for muta­ niles are better off ensconcing in a burrow and tion and selection to bring about genetic al­ diapausing, and subsequently remaining in the terations that compensate very recent changes same spot as an adult; most early summer in the environment. The available evidence juveniles are better off maturing directly and regarding cricket polymorphisms supports this emigrating; and every individual is therefore generalization. Thus, I do not find reasons best off, because of the bivoltine life cycle, if to support the frequently stated idea that (I) it has the potential for both capabilities and emigration or (2) suppression of reproductive has their expression adjusted by day length. tendencies have the evolved function of re­ lieving population pressure-i.e., they are sui­ cidal in nature. Instead, the evidence indicates G. veletis and G. pennsylvanicus that, except for incidental effects such as patho­ logical conditions arising from overcrowding, Post-diapause juveniles of veletis and early individuals suppress their reproductive tenden­ jmeniles of pennsylvanicus, present together cies or migrate during times of high popula­ in early summer, are both subjected to sharply tion density or habitat deterioration-in other increasing and long day lengths (14 hours in words when such behavior is most likely to mid-April to 16 + hours in mid-June at 45°- lead to increases in the realized reproduction 500 N). On the other hand, early (pre-dia­ of the individuals doing the suppressing. pause) juveniles of veletis and late juveniles The following quotations from the writings of pennsylvanicus, present together in late sum­ of contemporary biologists illustrate the contrary mer, are both subjected to sharply decreasing viewpoint: "The transformation of grasshoppers and short day lengths (16 + hours at July I into phase gregaria is a classical reaction of the to IO hours in late October at 45°-50° N). If sort referred to as 'density-dependent' in popu­ these species behave as do those others which lation ecology. Its function is clearly to produce have been tested (Masaki and Oyama, 1963; emigration that will relieve pressure on the soli­ taria foci, which are the most favorable breeding Fuzeau-Braesch, 1963, 1966), presumably the areas for the species" [Wilson, 1963]. "The major conditions to which the above stages are sub­ purpose of the emigration [of locusts] at this stage jected should accelerate development in all appears to be the relief of congestion, though, as a cases but one: shortening days, while accelerat­ minor or secondary advantage, it no doubt helps ing development of pre-diapause veletis juve­ the species to disseminate itself and colonize new niles, should decelerate or stop development of habitats as these become available" [p. 273, diapause-age veletis juveniles. Photoperiod may Wynne-Edwards, 1962]. MARCH 1968) LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 27

Fig. 11 summarizes the effects of several di­ consistently results in disappearance of the morphic or polymorphic tendencies known to ability to emigrate. Presumably, proportions of be related to population density in crickets. emigrating individuals should decrease, at a rate In general, high density populations result in influenced by the amount of immigration, for (1) higher tolerance of crowding, and (2) in­ the duration of each habitat-deme relationship creased emigration (dispersal). Both effects Each deme established by macropterous emi­ have clear probable advantages to individuals grants should initially have a high proportion possessing the dual potential. In a crowded of emigrants, and each disappearance of a situation, hyperaggressiveness except just fol­ habitat and extinction of its non-emigrant popu­ lowing mating is likely to diminish an indi­ lation would raise the proportion of emigrants vidual's mating success, and territorial be­ in the species. Density-related emigration can havior is futile. Thus, pair formation occurs be advantageous even where habitats do not more frequently by chance, and, as this ten­ disappear if differences in probability of breed­ dency increases, mating success becomes more ing success correlate closely with density, and closely linked with tolerating the presence of population densities fluctuate sufficiently among other individuals, and joining groups of tol­ demes to give emigrants a high probability of erant individuals. Such toleration by males entering low density demes. Otherwise, low should be minimized chiefly in connection with proportions of emigrants or deleterious emigra­ monopolization of a female, which is the one tion may persist because further loss of emigrat­ circumstance during which aggressiveness re­ ing ability is retarded by immigration, habitat mains high in crowded situations: a male is impermanence involving relatively long cycles, always hyperaggressive during a bout of cop­ or some combination of these events. ulations, driving other males from the vicinity Loss of emigrating ability in permanent habi­ (Alexander, 1961). tats can occur whether or not the difference In contrast, during periods of low density, between flying and non-flying individuals is aggressiveness and territoriality increase a genetic to begin with; when it is not genetic, male's chances of attracting and monopolizing any mutation eliminating the plasticity that females, since an increased proportion of leads to production of some fliers under some matings depend upon attraction of the females conditions, or reducing it, will be favored. On to individual calling males. A male with a the other hand, if emigration were suicidal as burrow is a more persistent, stationary caller, an evolved function, or at least had its chief and an aggressive male is more likely to retain effects on the individuals staying at home his burrow. In turn, possession of the burrow ("relieving population pressure"), we should enhances dominance (Alexander, 1961). expect dimorphism in regard to it to be retained A given individual's contribution to subse­ not only in normally nomadic species (some­ quent generations will be minimized as pop­ times called "fugitive" species), but also in non­ ulation size increases. This shift reduces the nomadic species living in permanent habitats relative risk taken by an emigrant, assuming in regions where a new invadable habitat does that colonization or the joining of other, less not appear and successful colonization cannot dense populations is possible. Emigration dur­ occur. ing times of high population density may The close relationship between photoperiod indeed increase the reproductive possibilities changes and flying ability in bivoltine crickets of individuals staying at home, but it is difficult such as Gryllus rubens and G. integer, discussed to view such an outcome as anything other than in a previous section, is particularly interesting an incidental effect of the selective advantage in these connections. Late summer is evidently accruing to emigrants under such conditions; a better time for this nomadic species to colonize further, there seems to be no reason for so than is early summer. A correlation with season considering it. as such may exist, for example, in availability In various Orthoptera, the close correlation of food for developing young. Another possi­ between habitat permanence, absence or scar­ bility is that a genetic factor has been incorpo­ city of invadable habitat, and loss of flying rated which enables individuals to predict the ability indicates that repeated loss of emigrants consistently much higher population density of 28 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43 late summer, using photoperiod as the cue. In humidity are relatively constant, and that they other words, selection may have favored a wing continue their 24-hour cycle for at least several length response based on photoperiod because days in continuous darkness. When light-dark photoperiod is a reliable indicator of other cycles were reversed, new peaks of activity ap­ seasonal changes that result in the second adult peared during the new dark period and eventu­ population of the summer being much larger ally replaced the old peaks. Continuous dark­ than the first one. Individuals are accordingly ness after reversal led to partial resumption of able to emigrate when emigration is the favored the original cycle in two males that had been alternative because of changes in population reversed four and five days. density, but they use photoperiod and not pop­ Nowosielski and Patton (1963, 1964) and ulation density as the cue. Mortality during Nowosielski, Patton, and Naegele (1964) studied winter is undoubtedly one of the reasons for the house cricket further, confirming nocturnal the smaller early summer population, but addi­ peaks of locomotory activity in both juveniles tional factors may be involved. and adults, but indicating late afternoon peaks of feeding, drinking, and courting behavior, EFFECTS OF WEA TiiER ON with high levels of these activities continued DAILY CYCLES OF ACTIVITY through most of the night. Rhythms of locomo­ tory activity persisted in continuous light or Most crickets, tropical or temperate, are noc­ dark and could be reset by changing the phase turnal, calling only or chiefly at night. Charac­ of the light-dark cycle; the "free-running" period teristically, species that live only in trees are was a little over 23 hours in continuous light, more strictly nocturnal than their close relatives just under 25 hours in continuous dark. living in weeds and grasses: for example, Cloudsley-Thompson (1958) reported that Oecanthus pini Beutenmuller and 0. laricis three adult females of the northern European Walker (Oecanthinae) living on pine and tama­ field cricket, Gryllus campestris, were most ac­ rack trees, respectively, rarely call during day­ tive in the daytime in the laboratory, with a light hours, while the four other North Ameri­ peak of activity about 3 pm. This pattern of can members of the Oecanthus nigricornis group activity persisted in continuous darkness and live on herbaceous vegetation and sing in large could be reset by (1) changing the phase of the numbers most of the day as well as at night. light-dark cycle; (2) providing two-hour light Variations in diurnal cycle also occur among periods after the rhythm had been damped by surface dwelling crickets. For instance, ground continuous darkness; (3) returning the environ­ and field crickets (Nemobiinae and Gryllinae) ment to 30°C after periods at lower tempera­ that live in the woods are usually more noc­ tures; and (4) supplying water after several days turnal than their relatives living in grasslands without water. or open fields. Alexander and Meral (1967) recorded the daily Some of these variations seem most likely to cycles of chirping of another northern cricket, be correlated with intensity of predation by Gryllus veletis, in southeastern Michigan by birds, but a few are not easily explained on that basis. Gryllus veletis and G. pennsylvani­ counting numbers of chirping males and found cus, for example, call both day and night east a striking difference on different years, evidently of the Rocky Mountains and can be heard at owing to differences in the general temperature any time of day continuously along hundreds regimes of the two years (Figs. 13-14). In 1965, of miles of grassy roadsides in the Great Plains. during the recording period and for some days Farther west the same species rarely chirp except before, the air temperature dropped too low for at night. chirping (about 50°F) shortly after sundown and Daily activity cycles have been studied experi­ rarely was too high (about 100°F) in chirping mentally in four crickets. Lutz (1932) used a locations during the day; during the same treadle arrangement to demonstrate that juve­ period in 1966, the night temperature did not niles and adult males of Acheta domesticus and fall below 50°F and the day temperature was Gryllus pennsylvanicus locomote more at night higher, probably too high for chirping in many than during the day when temperature and or most chirping locations. As a result, in l 965 MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 29

.... 1965 ~ 90 c June 25 •-·"'-. June 26 ·-·-·-• ~ 70 ./ • ./ ~ ~. / ~ 50 ,. • 0 30L______""-...... ···· _

25 June26

O'> 20 /\ June25 ·a.C -~ .c 15 () !\ . ./~ June 24 ;\ I I 1 (/) ~ \1 ; (I) : \/\\/ /\ . 10 • I • • I \ I .0 /I I I / I / E 1\ ._! \ I • J J \ / I I \ / ::::, • , , June 22 , \ , '. , z 1 5 ',J •-•,' June 23 I .! ---'d \ \ 1 \ 1\ •\ I • \ : \ // \ ,,./ \ ' •,.✓ ·-·, _:-=-~=::::--a--___. \ __------! 0 6 12 18 24 6 12 18 24 6 12 Hours of the day

:FIG.13. DAILY CHANGES IN NUMBER OF CHIRPING MALES OF Gryllus veletis DURING A 1965 PERIOD WHEN THE NIGHTS WERE Too COLD FOR CHIRPING BETWEEN SHORTLY AFTER SUNSET AND SHORTLY AFTER SUNRISE, AND WHEN No PART OF nm DAY WAS Too HOT FOR CHIRPING Temperature was not recorded June 22-24, hut the nights were cold, and there was rain on the night of June 22. veletis was strictly a daytime chirper, at least LIFE CYCLE CHANGES AND SPECIATION during the periods of observation, with the prin­ Bigelow (1958) and Alexander and Bigelow cipal chirping peaks in mid-morning and a les­ (1960) postulated that Gryllus veletis and G. ser one in mid-afternoon. In 1966, the same pennsylvanicus may have originated when a species in the same field chirped chiefly at night, tropical species developed both egg-overwinter­ though there were small daytime peaks at almost ing and juvenile-overwintering populations along the same times of day as in 1965. its northern border. Masaki (1961) described a These observations indicate that Gryllus vele­ juvenile-overwintering population in the south­ tis., and perhaps other field crickets living at ern part of the range of the otherwise egg­ extreme latitudes, can be either nocturnal or diapausing species, Scapsipedus aspersus, sym­ diurnal, depending upon the prevailing weather patric with the parental population but with conditions in any given year in the variable the adults of the two populations seasonally climates where they occur. How quickly such separated. Earlier in this paper, I described a reversals can take place, precisely how they oc­ temporally isolated, fall-adult population in the cur, and whether or not weather conditions southern part of the range of G. fultoni, a spe­ during juvenile life affect the daily activity cycles cies which farther north overwinters only in the of adults are questions that remain to be juvenile stage and is adult only in the spring. answered. Fulton ( 1951) and Alexander ( 1957) described 30 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43

1966 ·a,.... .c 90 ...... •-·-•-• ...·-·,. .--•-•'• ...... -•-•-- ...... C ...... ,,,..-. '· / ...... __. ~ 70 .c -·-·-·-...... --· / "·---•---•--- . 0 LL 50 0 30

15 Ol ·e-C ·,/·,/·\ :.c 10 0 ("/'\ Ju:25 / • • /\ June26 / ~ 0 Q.) .0 5 E i ·-·\ r \ tf ::::, ·-· . •-· ·-•/ z 01------.-----,-----.----,------,.-----,-----.--'"---..---'·-·/ '·,/ IB 24 6 12 18 24 6 12 18 Hours of the day

FIG. 14. DAILY CHANGES IN NUMBER OF CHIRPING MALES OF Gryllm veletis DURING A 1966 PERIOD WHEN THE NIGHTS WERE NEVER Too COLD FOR CHIRPING, AND WHEN MoST LOCATIONS ,VERE Too HOT FOR CHIRPING DURING SEVERAL HOURS OF MIDDAY

temporally isolated, spring-adult populations in is it necessary or reasonable to postulate that the northern part of the range of G. firmus. geographic isolation was in some way involved None of these temporally isolated populations in the development of life cycle differences is geographically isolated from the rest of the between closely related species or populations species. of the same species? Alexander and Bigelow These four cases, and the hypotheses given (1960), in referring to G. veletis and G. penn­ earlier that relate to them, all imply a connec­ sylvanicus, maintain (p. 344): "It is doubtful tion between the appearance of a new life cycle that the possibility of allopatric speciation could and the multiplication of species. In fact, the ever he completely ruled out, but the evidence presence of sibling species with radically dif­ suggests that geographic isolation is both un­ ferent life cycles, such as G. veletis and G. penn­ likely and unnecessary in this particular case." sylvanicus, is enough to raise the question of There is a remarkable similarity between this whether the life cycle change contributed to the statement and the following one made by Ghent speciation process or was merely incidental to and Wallace (1958) (and not seen by me until it. This question is particularly interesting with 1965) in trying to explain a similar case of egg crickets and their relatives (1) because both of and pre-pupal overwintering with seasonal sep­ the common temperate life cycles discussed aration of adults in Neodiprion sawflies: "The earlier occur within species as well as between change in overwintering habit effects temporal siblings; (2) because no genetic difference is isolation of breeding adults with complete effi­ necessary for essentially complete temporal isola­ ciency, and as a pathway of speciation, it calls tion of the relatively short-lived adults (Figs. more urgently for a premise of geographical con­ 2, II); and (3) because no case is known of tinuity than for one of geographical isolation" spatial isolates with the two different life cycles, p. 271. Mayr (1963a), on the other hand, believed either sibling species or conspecific po_pulations. (p. 477): "It would be altogether unlikely for The next question seems to be: to what extent a population to adapt itself simultaneously to MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 31 two different breeding cycles. Where a species shift the life cycle, though it might ad just evolves two such cycles it does so in different, slightly the timing of the adult periods. The geographically segregated, populations ... " cycle itself is evidently restricted by possible [and] "It is rather probable that the difference overwintering stages; there are no known inter­ in breeding cycle of [G.] pennsylvanicus and mediate overwintering stages between eggs and [G.] veletis was not yet complete, when they late juveniles and, therefore, no seasonally in­ first met, after emerging from their geographic termediate periods of adulthood. The only pos­ isolation. If so, competition eliminated any sible "intermediacy" is represented by late pennsylvanicus tendencies for spring breeding in adults of veletis or early adults of pennsylvani­ and of fall breeding in veletis." Alexander (1963) cus,· but veletis and pennsylvanicus currently discounted the likelihood of Mayr's interpreta­ have about the same adult periods as other uni­ tion, pointing out that certain species have in voltine crickets that overwinter respectively as fact evolved two cycles in one locality, but Mayr juveniles and eggs and lack seasonally different (1963b) responded (p. 206): "To me, at least, it sympatric siblings (Figs. 2, II). No evidence would be far simpler to have the ancestral spe­ exists, therefore, that competition between cies of the veletis-pennsylvanicus group split veletis and pennsylvanicus has materially af­ into at least two border populations, one let us fected their life cycles. say east and another one west of the Appala­ Because of the history of the discussion about chians, furthermore, to assume that in one of the significance of seasonal isolation in specia­ these egg-overwintering was more advantageous, tion, and the general attitude among biologists in the other one, juvenile-overwintering, and toward the possibility of speciation without geo­ that with the post-Pleistocene improvement of graphic separation, it seems useful to account the climate, the two populations could overlap and coexist, owing to their previously estab­ for other disagreements between Mayr (1963a, lished seasonal separation." Mayr concluded by 1963b) and Alexander and Bigelow (1960) and repeating his earlier conclusion that "in not a Alexander (1963). single case is the sympatric model superior to Mayr (1963b) (I) says that it is not stated how an explanation of the same natural phenome­ the northern and southern populations of a non through geographic speciation." species such as the ancestor of G. firmus might Mayr's (1963a) postulate that seasonal separa­ have been geographically separated so that a tion was completed because of competition fol­ process of allochronic speciation could occur lowing establishment of sympatry brings up within the northern segment; (2) believes that some points evidently requiring clarification. not one but a series of peripheral isolates would Two kinds of competition could be meant: (1) have been involved; (3) suggests that there may between adults at mid season, particularly in have been an east-west split between two such connection with reproductive activities, and (2) border populations (contrasting this with a between juveniles of the two populations, or north-south split); and (4) believes that one between juveniles of one and adults of the isolate became solely juvenile-overwintering and other, particularly in connection with food. It the other solely egg-overwintering, and that is difficult to imagine how either of these kinds speciation occurred allopatrically between them. of competition could be eliminated by shifts in Concerning the first point, geographically life cycle. First, adult and juvenile crickets (pos­ isolated populations of G. firmus in the inland sibly excepting very tiny juveniles) live in the sand hills of North Carolina were first described same places and eat the same things. Competi­ by Fulton (1952), and later discussed by tion of the second kind, therefore, probably still Alexander (1957), who noted that they hy­ exists between G. veletis and G. pennsylvanicus, bridized with coastal populations (unpubl. tests and seemingly could not be changed by any by Fulton) and had the same song. Additional possible shift in their life cycle. Second, compe­ inland populations in Alabama have been found tition (or interference) between adults of the since then (Fig. 4), some of which contain indi­ two species also could not easily be construed to viduals with different life cycles. The example 32 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43 of a north-south split is, therefore, already before TABLE 3 us in the proper species group, the right geo­ Successes and failures in 16 attempts to hybridize graphic locations, and (at least today) the right field cricket (Gryllus) species with known diapause climate. stages• Secondly, the number of successful peripheral isolates does not seem to be important in assess­ MATING OFFSPRING ing the significance of seasonal isolation in caus­ A. Non-dia{>ausersX non-diapausers ing speciation. Actually, the more isolates there I. ass1milis X bermudensis + are, the more likely it is that allochrony was in­ B. Non-diapausers X juvenile-diapaus- volved in this case, and the less likely that al­ ers lopatry was more than incidental to the specia­ I. bimaculatus X campestris + 2, bermudensis X campestris + tion process. The reason for this assertion is the 3. assimilis X rubens + following: if a species tends to produce popula­ 4. assimilis X integer + tions with two kinds of life cycles, as, for 5. assimilis X veletis example, G. firmus does, then an increased C. Non-diapausers X egg-diapausers I. assimilis X pennsylvanicus + number of isolates will mean it is less likely D. Juvenile-diapausers X juvenile-dia- that none contains both life cycles. The uni­ pausers versal success of breeding tests among related I. veletis X veletis + Gryllus species that overwinter in the same stage 2. veletis X rubens + 3. veletis X fultoni + (Table 3), and the absence today of song or 4. veletis X vernalis + ecological (habitat) differences between veletis 5. vernalis X fultoni + and pennsylvanicus, indicate that any persistent, E. Egg-diapausers X egg-diapausers double-cycled, ancestral population would have l. firmus X pennsylvanicus + contributed genes to both incipient species. Ac­ F. Egg-diapausers X juvenile-diapausers 1. pennsylvanicus X veletis cordingly, the significant isolation would still 2. pennsylvanicus X rubens have been temporal from the start, and what­ ,.., 3. firmus X veletis ever geographic isolation occurred would 4. firmus X rubens have been incidental. Thus, speciation would •, summarizing Fulton (1951), Alexander (1957, and have occurred in the north even if every unpublished 1967), Bigelow (1958, 1962), and Cousin isolate had been double-cycled. Furthermore, (1961). unless the temporal isolation were effective as +. offspring obtained in at least some attempts. an extrinsic separator (in fact, capable of caus­ -, no offspring obtained. .. , from Alexander, 1967 (unpublished). ing speciation), double-cycled populations would Only species for which diapause seems well under­ tend to destroy the effects of prior geographic stood are included. The most extensive attempts have been carried out between veletis and pennsyl­ separation when they became sympatric with vanicus and between veletis and firmus. geographic isolates (incipient species) that had developed different, single cycles while living "east-west" split is known to have occurred in in different places. In other words, even if G. firmus (Fig. 4). Mayr's postulate of an east­ there were evidence that geographic isolates west split on either side of the Appalachians sometimes develop different cycles (and there is not unreasonable, but it somewhat confuses is none), we are not free to consider the fact the issue since the only important point is that by itself until we can eliminate the likelihood the populations involved must be separated in that double-cycled populations will also persist. some fashion from the continuously breeding Thirdly, whether the split, or separation of populations farther south in Florida; otherwise double-cycled northern populations from south­ the latter would provide a means of indirect ern non-cycled populations, is "east-west" or gene exchange between sympatric northern "north-south" does not matter as long as it took spring-maturing and fall-maturing populations. place somewhere in the northern part of the This required break has no relevance to the range of the species - or in the region where question of geographic isolation between the there are mild winters, as there are today in the two northern populations, which need never be sand hill region of North Carolina where an allopatric during their speciation; three species, MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 33 not two, would result from the one geographic we must wonder why, in species like firmus discontinuity. Likewise, the particular timing and fultoni, populations with one of the two of the geographic break relative to the genetic possible cycles are inevitably small, sympatric divergence of the sympatric northern popula­ portions of a large population with the other tions from one another is of no relevance. cycle; likewise veletis has never been found Mayr's fourth point of concern - whether life without pennsylvanicus, though pennsylvanicus cycle differences now existing between sibling occurs without veletis in Nova Scotia. species arose during allopatry or during sym­ The second possibility above, a sudden shift patry - is evidently the critical one. in one generation of an entire population from First we should check natural populations for one life cycle to another, seems too far-fetched evidence. The four cases I have already men­ to consider seriously. Only slightly more likely tioned are pertinent: Gryllus firmus, G. fultoni, is the third possibility that some individuals G. carnpestris, and Scapsipedus aspersus. In all would shift cycles and those with the old cycle four cases populations with the two kinds of would subsequently disappear leaving only the life cycles coexist in the same regions, in the new cycle. But if we must involve allopatry this southern part of the range of the three more third possibility seems to be the only one. To or less northern species and in the northern propose allopatric development of life cycle part of the range of G. firmus, the only southern differences when sympatric duality in life cycles species. Recourse to natural situations, there­ is known is to require a more complex hy­ fore, does not support Mayr's argument, for in pothesis than is necessary to take account of the no cricket, and to my knowledge in no other facts. arthropod, are there allopatric populations with Suppose allopatry is not required. Can the seasonally staggered life cycles correlated with pairs of seasonally different populations that different overwintering stages. persist today in a completely sympatric relation­ This finding is no surprise. The kind of life ship with members of the larger (parent) popula­ cycle shift it would involve can occur in three tion that still possesses the original cycle, fail possible ways: (l) a continuously breeding pop­ to diverge, since they overwinter in different ulation moving north in two or more locations stages and are subjected to different tempera­ could develop one kind of life cycle in one ture and photoperiod regimes during the vari­ place and a second kind in another place; (2) ous parts of their life cycle? I suggest that selec­ one or more populations already having a single tion would strongly favor divergence between life cycle could make a complete shift in one such populations from the start, particularly generation to the other cycle, or (3) one or more in regard to overwintering behavior and physi­ single-cycled populations could develop a ology. We have already seen (Table 2) what double life cycle and then lose the original striking effects alterations of life cycle timing cycle through extinction (shift slowly from one can have, in the absence of any genetic altera­ cycle to the other). tion, in the field performance of species such No case is known of the first of these three as Gryllus rubens, G. integer, and G. firmus, and possibilities, which has already been discussed. in laboratory populations of G. campestris, But if it is used to account for the origin of Nemobius yezoensis, and others. Even if some species such as G. veletis and G. pennsylvanicus, gene flow occurred at mid-season between sym­ two problems are introduced. First, if a third patric, differently cycled populations, or because (southern) derived species is involved, such as of repeated shifting back and forth between the G. firmus in this case, then the near-identity of two life cycles, divergence would surely con­ veletis and pennsylvanicus in a variety of re­ tinue if only because of the selective advantage gards in which they differ from firrnus, such as in the acquisition of different kinds of special chirp rate and pulse rate in calling, number of overwintering characteristics, owing to the dif­ file teeth on the tegmina, and general size and ference in the life stage that repeatedly enters body form, must be accounted for by parallel winter. evolution or divergence only by firmus from the One question that might be raised, and common ancestor of all three species. Second, perhaps was involved in Mayr's objections, 34 THE QUARTERLY REVIEW OF BIOLOGY (VOLUME 43 concerns the problem of survival of newly off­ appropriate diapause rather quickly, any strong cycle individuals in the face of competition reduction of gene flow between populations from the established population. Lewis (1966), with different over-wintering stages could, there­ for example, raises this question in attempting fore, cause an irrevocable shift toward total to describe how chromosomal reorganization inter-sterility. This kind of divergence seems may result in speciation without geographic appropriately described in crickets as a "short isolation. That this problem is probably not route" to speciation, for total sterility is actually pertinent to allochronic speciation in crickets rare among sibling species of most or all kinds is indicated by the coexistence of seasonally of animals. But early inter-sterility is not a different populations within species. required part of the process of speciation by If no gene flow at all occurs between season­ seasonal separation as outlined here. ally cycled populations, then speciation would Both spring and fall adults of firmus have occur almost exactly as if the separation were repeatedly produced hybrids in matings with geographic. If gene flow persists, however, then pennsylvanicus, but neither has hybridized with divergent selective action (roughly equivalent veletis. This may be an indication of ancestral to a combination of disruptive selection and northern affinities in firmus, and along with the partial extrinsic separation) will result in specia­ more northward extension and greater northern tion only if overwintering differences in some abundance of egg-overwintering populations of way eventually become linked to incompati­ firmus (Fig. 4), suggests a close relationship with bility. pennsylvanicus._ Because of this information, the Here an unexplained phenomenon enters the possibility cannot be eliminated that firmus is picture, at least with regard to veletis and penn­ a derivative of a pennsylvanicus-like ancestor, sylvanicus. Among some 40 attempts to hy­ rather than vice versa. This would mean that, bridize different cricket species, of which about if present distribution and sterility barriers are half have been really intensive efforts, four of taken into consideration, spring-adult firmus the five complete failures known to me have may be independently derived (Figs. 15, 16). It involved northern crickets with different over­ could also mean that veletis and pennsylvanicus wintering (diapause) stages (Table 3). No hy­ may have diverged by production of a double­ brids have even been produced between species cycled population in the southern part of the with different diapause stages, even though three range of a northern species rather than vice investigators in North America (Fulton, 1952; versa. This question affects the relative prob­ Bigelow, 1958, 1960; Alexander, 1957, and un­ ability of the two models for allochronic spe­ publ.) have tried several cases. ciation illustrated in Fig. 15 explaining the origin of veletis and pennsylvanicus, but it does The closest case is Masaki and Ohmachi's (1967) successin hybridizing an egg-diapauser with what not reduce the likelihood that seasonal separa­ they call a "facultative" juvenile-diapauser (depen­ tion was the cause of speciation. dent upon photoperiod). Their results are rather In summary, it would seem difficult to pre­ complex and interesting: for example, some female vent genetic divergence between persisting F1 hybirds took longer to mature without any seasonally different populations, even if they delay in egg hatching time than even the faculta­ were sympatric, even if some gene flow persisted, tively juvenile-diapausing parent, and the recipro­ and even if hybrids were not adult at the cal crosses gave quite different results. wrong season and did not enter winter during This failure of hybridization between species other than the two winter-hardy stages. The with different diapause stages is almost the only eventual result of such divergence would be generalization about incompatibility between two species that overwintered in different stages Gryllus species that can be presented at this and were adult at different seasons, but had time. Both copulation and oviposition (less fre­ essentially the same habitats and occupied, at quent in the interspecific matings) occurred in least initially, the southern portion of the all such crosses tried by me, but no eggs hatched. parental range in the case of a northern species Since populations of crickets consistently enter­ and the northern portion of the parental range ing winter in different stages ought to develop in the case of a southern species. MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 35

seasonal Isolation Without Geographic Separation

North Example Egg Diapause Gryllus :TI17 "'ff' \ I pennsylvanicus •1r•Eoo-, SPRINGADULTS FALL ADULTS

SPRINGADULTS FALL ADULTS Juvenile-overwintering\ Egg- overwinteringI

Wintersallow only two overwinteringI stages Day~ lenglhsana tempera!uresdo not South causeall indMduols to d,apause TEMPERATESPECIES 'TROPICAL SPECIES -(or climates changing)- MOVINGSOUTH MOVINGNORTH

FIG. 15. ALTERNATIVE HYPOTHESES FOR THE MULTIPLICATION OF FIELD CRICKET SPECIES WITHOUT THE NECESSITY OF GEOGRAPHIC ISOLATION Left, (a) a northern, egg-diapausing cricket producing spring populations in the southern part of its range; (b) a northern, juvenile-dil\p;msing cricket producing fall adults in the southern part of its range. Right, a southern cricket devcµoping both of the prevalent temperate life cycles in the northern part of its range. Black arrows represent parts of the diagrams that have been demonstrated in natural populations; partly black arrows indicate possible situations in known species; white (open) arrows indicate hypothesis.

There does not seem to be any easy way to (l) Multiple-year life cycles and univoltine derive veletis and pennsylvanicus except by one life cycles with short-lived adults are widespread of the two possible variations of this scheme, among temperate insects (Pickford, 1953; which Alexander and Bigelow (1960) termed Brooks, 1958; Alexander and Moore, 1962). "allochronic" speciation in reference to the Both involve potential temporal isolation of central role of extrinsic temporal separation. adults. As Walker (1964) points out, there is "Allohoric" speciation is a more specific term less expectation of rapid divergence in the case referring directly to seasonal sep;p-ation. of multiple-year life cycles; geographic restric­ It seems impossible even to ga!n an inkling tion of populations emerging in different years concerning how frequently allochronic specia­ is also more likely, and is known in various tion may have occurred in various kinds of cicadas (Alexander and Moore, 1962). insects until intensive systematic-biological work (2) Unexplained variations in overwintering has been carried out on a large number of stage are prevalent in all major groups of tem­ temperate insect groups by investigators who perate insects, both within and between species are at least aware of the possibility and know as presently recognized. how the process can work. Such studies ob­ viously have not yet been accomplished, for [Any entomological text or periodical in which the possibility had not been considered prior life histories are described can be used to verify this remark. Metcalf, Flint, and Metcalf (1951) to presentation of this case and that of Ghent is an excellent example, and their remarks on and Wallace (1958). For at least three reasons the life history of "the field cricket" in the I believe a large number of additional cases United States, ".1.1.cheta assimilis Fabricius" may be discovered: (p. 528-529), constitute an instructive illustra- 36 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME 43

univoltine bivoltine breed all year egg diapause juvenile diapause assimilis bermudiensis firmus pennsy/vanicus veletis verna/is fultoni rubens integer (90} (145) (17) (27) (27) (31) (44) (57) (77)

xenc I mesic (and I hill prairies) ~sand beaches ...._? grasslands

ORIGIN TEMPERATEORIGIN ?

FIG, 16. A DIAGRAM OF THE PROBABLE PHYLOGENETIC RELATIONSHIPS OF THE FIELD CRICKETS OF EASTERN NORTH AMERICA, INCLUDING THEIR PROBABLE GEOGRAPHIC ORIGINS, HABITAT SHIFTS AND LIFE CYCLE CHANGES Although firmus seems everywhere to have a tendencr. toward egg diapause, some northern populations overwinter as juveniles, with no evidence of juvenile diapause (see text for further explanation). Numbers in parentheses below species names are approximate wingstroke rates in the calling song at 80° F; rubens and integer are trillers, assimilis a B. chirper (Alexander, 1962), fultoni a Ba chirper, vernalis intermediate between Bi and B.,; all the rest are Bi chirpers. All morphological characters studied su_pport this arrangement (see Alexander, 1957; Alexander and Walker, 1962, for all species but bermudens1s and integer, which are morphologically most like firmus and rubens, respectively).

tion: "The life history varies greatly throughout they lead either to multiple-year life cycles or the range of the insect, but in the northern to seasonal separation of adults. As an example states the majority hibernate in the egg stage of the possibilities, Brooks (1958) has presented and complete a single generation in a year, evidence that populations of a grasshopper, while in the Imperial Valley of California and Arphia conspersa Scudder, in Alberta, Canada, the Gulf Coast they winter as nymphs or remain more or less active the year round and may pro­ contain some individuals that mature in two duce as many as three generations .... Through· years, overwintering first as eggs and then as out most of the northern states, adults are pres­ juveniles and producing spring adults, and ent from late July until heavy frosts and lay other individuals that mature in one year, over­ their eggs mostly in August and early Septem­ wintering only as eggs and producing fall adults. ber." This attempt to generalize around an im­ (3) Cryptic species are abundant in every possible taxonomic situation resulted in a series well-studied insect group, and most insect groups of remarks in which scarcely a single clause is (tens of thousands!) have not yet been well correct.] enough studied to detect them. Every life cycle Except for species with many generations per change that occurred within a temperate region year, or long-lived adults that reproduce during could have involved allochronic speciation, and successive seasons or cause overlap of genera­ every pair of sibling species with different adult tions, when such variations are intraspecific seasons and overwintering stages may properly MARCH 1968] LIFE CYCLE ORIGINS AND SPEC/AT/ON IN CRICKETS 37 be suspected of having speciated allochronically. and Danilevskii, 1965). Relatively few studies, Even geographic speciation can only be in­ however, deal with life cycle variations within ferred in the vast majority of cases. As a conse­ narrow taxonomic groups, their significance in quence there will never be a successful argu­ relation to differences in habitats and geo­ ment against those who insist that no alterna­ graphic ranges, and the problems in accounting tive is possible. It does not matter whether for their evolutionary origins. Ohmachi and this insistence takes the form of a flat statement Matsuura (1951), Fulton (1952), Walker (1963), or simply the rejection of all alternative hy­ Masaki and Ohmachi (1967), Pickford (1953), potheses through construction of a geographic and Brooks (1958) are notable exceptions, with one that is continually altered and defended as discussions of life cycle variations among closely more likely in the face of contrary evidence. Spe­ related species of crickets and grasshoppers. ciation by other than geographic isolation may Danilevskii (1965) has reviewed other studies of justifiably be inferred whenever sibling species interest in this regard. However, Ghent and are not geographically related in the ways com­ Wallace (1958), Bigelow (1958), Alexander and mon to sibling species, and when they simul­ Bigelow (1960), Masaki (1961), and Masaki and taneously exhibit a striking ecological difference Ohmachi (1967) seem to be the only investiga­ (such as a host difference or a life history dif­ tors who have seriously endeavored to account ference) that effectively separates their adults for life cycle differences between closely related and does not appear to have arisen through species. As already indicated, the conclusions of competition or other interaction between the Ghent and Wallace (1958) and Masaki (1961) species. Such facts are not only inconsistent with are strikingly similar to those of Bigelow (1958) a geographic hypothesis, they may cause the and Alexander and Bigelow (1960), discussed necessity of accounting for a geographic differ­ and enlarged by Alexander (1963, and in the ence to become an onerous complication. If, present paper). under the above conditions, an alternative to It would seem that, except for analyses toward geographic speciation exists that is reasonable understanding of certain physiological mecha­ in its totality, based on the characteristics of nisms in a relatively few cases, we have really the organisms under consideration, and if inter­ made only the barest start toward understanding mediate steps in such a presumed alternative the morphological, physiological, and behavioral process are found in natural populations, geo­ complexities that selection has produced in dif­ graphic speciation may become an extremely un­ ferent insects that live in different climates. likely possibility. Reasonable progress toward generalizations In view of the importance of explaining rates concerning adaptation to climate seems possible and distribution of speciation, I think it will be only if we keep in mind, on a larger scale than unfortunate if attention to extrinsic temporal has been the case, that appropriate comparisons isolation is delayed by lumping it with some always involve not one but two kinds of vari­ of the more outlandish attempts to explain spe­ ables: (I) climate and habitat (environment) ciation without extrinsic isolation of any kind, and (2) phylogeny. We cannot hope to under­ or discouraged by pronouncements deriving stand the significance of similarities and differ­ from previous arguments over the old specter ences between species living in the same or of "sympatric speciation" in the sense of no different climates unless we know something extrinsic isolation. On the other hand, even if about the similarities and differences in the field crickets were so peculiar that allochronic equipment and way of life with which each speciation turned out to be trivially infrequent, began. I think we would still want to understand it in this case. ACKNOWLEDGMENTS I am indebted to Daniel Otte of The University CONCLUSIONS of Michigan for photographing Gryllus veletis (Fig. 7) and to Franz Huber of The University of The literature on diapause and related phe­ Munich (Germany) for allowing me to use his nomena in arthropods is massive (see reviews photographs and information concerning Gryllus by Umeya, 1950; Masaki, 1963; Browning, 1952; campestris (Fig. 8). Angus Hall and David Chesluk 38 THE QUARTERLY REVIEW OF BIOLOGY [VoI.UME43 served as faithful and perceptive research assistants Science Foundation Grant, GB-251. Data on the in 1964 and 1965, respectively; Chesluk carried out 1965 populations of veletis and pennsylvanicus were much of the work to obtain the data in Fig. 6. gathered and compiled by Gerald H. Meral (now Bertram G. Murray, Mary Jane West, Daniel a graduate student at The University of California, Otte, Theodore H. Hubbell, and Donald W. Tinkle, Berkeley) while he held an award from a National all of The University of Michigan, made helpful Science Foundation Undergraduate Research Par­ suggestions concerning the manuscript. Dr. and ticipation Grant to the Department of Zoology, The Mrs. Leonard Wing allowed me to use their field in Superior Township, Washtenaw County, Michigan, University of Michigan. Meral also made numerous during 1964-66, and kept it free from disturbances detailed observations throughout the 1965 season during my study. which have been incorporated into other portions Most of this work was supported by a National of the discussions.

LIST OF LITERATURE ALEXANDER,R. D. 1957. The taxonomy of the --. 1960. Interspecific hybrids and speciation field crickets of the eastern United States in the genus Acheta (Orthoptera, Gryllidae). (Orthoptera: Gryllidae: Achei'a). Ann. En­ Canad. ]. Zool., 38: 509-524. tomol. Soc. Amer., 50: 584-602. --. l 962. Factors affecting developmental rates --. 1961. Aggressiveness, territoriality, and sex­ and diapause in field crickets. Evolution, 16: ual behavior in field crickets (Orthoptera: 396--406. Gryllidae). Behaviour, 17: 130-223. BIGELOW,R. S., and P:S. A. CocHAUX. 1962. --. 1962. Evolutionary change in cricket acous­ lntersterility and diapause difference between tical communication. Evolution, 16: 443-467. geographical populations of Teleogryllus com­ --. 1963. Animal species, evolution, and geo­ modus (Walker) (Orthoptera: Gryllidae). graphic isolation. Syst. Zoo/., 12: 202-204. Austral. ]. Zool., IO: 360-366. --. 1964. The evolution of mating behavior BLATCHLEY,w. s. 120. Orthoptera of North­ in arthropods. Symp. Royal Entomol. Soc. eastern America. Nature Puhl. Co., Indianap­ (Lond.), 2: 78--94. olis. ALEXANDER,R. D., and R. s. BIGELOW.1960. BROOKS,A. R. 1958. Acridoidea of southern Al­ Allochronic speciation in field crickets, and a berta, Saskatchewan, and Manitoba (Orthop­ new species Acheta veletis. Evolution, 14: tera). Canad. Entomol. Suppl., 9: 1-92. 334-346. BROWNING,T. 0. 1952. The influence of tempera­ ALEXANDER,R. D., and G. MERAL. 1967. Seasonal ture on the rate of development of insects, and daily chirping cycles in the northern with special reference to the eggs of Gryllulus spring and fall field crickets, Gryllus veletis commodus Walker, Austral. ]. Sci. Res, B, 5: and G. pennsylvanicus. Ohio ]. Sci., 67: 200- 96-Ill. 209. BURMEISTER,H. 1838. Handbuch der Entomolo­ ALEXANDER,R. D., and T. E. MOORE. 1962. The gie. Enslin, Berlin. evolutionary relationships of 17-year and 13- CHAUVIN,R. 1958. L'action du groupement sur year cicadas, and three new species (Homoptera, la croissance des grillons (Gryllulus domesticus). Cicadidae, Magicicada). Univ. Mich. Miscell. ]. Insect Physiol., 2: 235-248. Puhl. No. 121, 59 p. ALEXANDER,R. D., and E. s. THOMAS. 1959. Sys­ CLOUDSLEY-THOMPSON,J. L. 1958. Studies in di­ tematic and behavioral studies on the crickets urnal rhythms. VIII. The endogenous chrono­ of the Nemobius fasciatus group (Orthoptera: meter in Gryllus campestris L. (Orthoptera: Gryllidae: Nemobiinae). Ann. Entomol. Soc. Gryllidae). ]. Insect Physiol., 2: 275-280. Amer., 52: 591-605. CousrN, G. 1961. Essai d'analyse de la speciation ALEXANDER,R. D. and T. J. WALKER. 1962. Two chez quelques gryllides du continent americain. introduced field crickets new to eastern United Bull. Biol. France-Belg., 95: 155-174. States (Orthoptera: Gryllidae). Ann. Ent. Soc. DANILEVSKII,A. S. 1965. Photoperiodism and Sea­ Amer. 55: 90-94. sonal Development of Insects. Oliver and .Boyd, ARNDT,C.H. 1931. The Haitian coffee tree cricket. Edinburgh. ]. Dept. Agric., Puerto Rico, 15: 325-335. ELLIS,P. E. 1959. Learning and social aggregation BIGELOW,R. S. 1958. Evolution in the field in locust hoppers. Anim, Behavior, 7: 91-106. cricket, Acheta assimilis Fab. Canad. ], Zool., FULTON,B. B. 1951. The seasonal succession of 36: 139-151. orthopteran stridulation near Raleigh, North MARCH1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 39

Carolina. J. Elisha Mitchell Sci. Soc., 67: HUBBELL,T. H. 1938. New cave-crickets from 87-95. Yucatan, with a review of the Pentacentrinae, --. 1952. Speciation in the field cricket. Evolu­ and studies on the genus Amphiacusta (Orthop­ tion, 6: 283-295. tera, Gryllidae). Carnegie Inst. Wash. Pub!., FuzEAU-BRAESCH,S. 1960. Etude biologique et bio­ 491: 191-233. chimique de la pigmentation d'un insecte LAURENTIAUX,D. 1951. La probleme des Blattes Gryllus bimaculatus De Geer (Gryllide, Orthop­ paleozoiques a ovipositeur exteme. Ann. tere). Bull. Biol. z,·r. et Belg., 105: 525-627. Paleont., 37: 185-196. --. 1961. Variations dans la longueur des ailes LF.Es,A. D. 1955. The Physiology of Diapause in en fonction de l'effet de groupe chez quelques Arthropods. Univ. Press, Cambridge. especes de Gryllides. Bull. Soc. Zool. Fr., 86: LEVITA,B. 1962. Contribution a l'etude du me­ 785-788. canisme d'un elfet de groupe chez un insecte --. 1962. A propos de la specificite des stimuli orthoptere: Gryllus bimaculatus De Geer. sensoriels chez !es insectes. Compt. Rend. Soc. Bull. Soc. Zool. Fr., 87: 197-221. Biol. (Paris), 156: 1577-1581. LEWIS, H. 1966. Speciation in flowering plants. --. 1963. Ajustement du cycle de vie avec Jes Science, 152: 167-172. saisons chez un insecte univoltin. Compt. LINDROTH,C. H. 1949. Die fennoskandischen Ca­ Rend. Acad. Sci. (Paris), 256: 792-794. rabidae. Eine tiergeographische Studie. Gote­ --. l 965. Hibernation de Gryllus campestris borgs Velensk. Samh. Hand/. (B), 4: 1-911. L. (Orthop.: Gryllides): Analyse de la stabilite LUTZ, F. E. 1932. Experiments with Orthoptera et des exigences de Ia diapause. Compt. Rend. concerning diurnal rhythm. Amer. Mus. Soc. Biol. (Paris), r59: 1048-1052. Novit., 550: 1-24. --. 1966. Etude de la diapause de Gryllus MASAKI,S. 1960. Thermal relations of diapause campestris (Orthoptera). J. Insect Physiol., 12: in the eggs of certain crickets (Orthoptera: 449-455. Gryllidae). Hirosaki Univ. Fae. Agric. Bull., GAnBurr, P. D. 1959. The bionomics of the wood 6: 5-20. cricket, Nemobius sylvestris (Bose) (Orthop­ --. 1961. Geographic variation of diapause in tera: Gryliidae). J. Anim. Ecol., 28: 15-42. insects. Hirosaki Univ. Fae. Agric. Bull., 7: GHENT,A. W., and D. R. WALLACE. 1958. Ovi­ 66-98. position behavior of the swaine jackpine sawfly. --. 1963. Adaptation to local climatic condi­ Forest Science, 4: 264.-272. tions in the Emma field cricket (Orthoptera: GHOSH,C. C. 1912. The big brown cricket. Mem. Gryllidae). Konty-a, 31: 249-260. Dept. Agric., India, 4: 161-182. --. 1965. Geographic variation in the intrinsic HAYSLIP,N. C. 1943. Notes on biological studies of incubation period: a physiological dine in the male crickets at Plant City, Florida. Fla. Emma field cricket (Orthoptera: Gryllidae: Entomol., 26: 33-46. Teleogryllus). Hirosaki Univ. Fae. Agric. Bull., HEBARD,M. 1934. The Dermaptera and Orthop­ II: 59-90. tera of Illinois. Ill. Nat. Hist. Surv. Bull., --. 1966. Photoperiodism and geographic varia­ 20: 125-279. tion in the nymphal growth of Teleogryllus HENNEGUY,L. F. 1904. Les Insectes. Morphologie, yezoemma (Ohmachi and Matsuura) (Orthop­ Reproduction, Embryogenie. Masson, Paris. tera: Gryllidae). Konty-a, 34: 277-288. HoGAN,T. W. 1960. The onset and duration of diapause in eggs of Acheta commodus (Walk.) --. 1967. Geographic variation and climatic (Orthoptera). Austral. ]. Biol. Sci., 13: I 4.-29. adaptation in the field cricket (Orthoptera: --. 1965. Some diapause characteristics and in­ Gryllidae). Evolution, 21: 725-741. terfertility of three geographic populations of .MASAKI,s., and F. OHMACHI: 1967. Divergence Teleogryllus commodus (Walk.) (Orthoptera: of photoperiodic response and hybrid develop­ Gryllidae). Austral. ]. Zoo/., 13: 455-459. ment in Teleogryllus (Orthoptera: Gryllidae). --. 1966. Physiological differences between races Konty-a, 35: 83-105. of Teleogryllus commodus (Walker) (Orthop­ MASAKI,S., and N. OYAMA. 1963. Photoperiodic tera, Gryllidae) related to a proposed genetic control of growth and wing-form in Nemobius approach to control. Austral. ]. Zoo/., 14: 245- yezoensis Shiraki (Orthoptera: Gryllidae). 251. Kontyu, 31: 16--26. --. 1967. Inter-racial mating of a non-diapaus­ MAYR. E. 1963a. Animal Species and Evolution. ing race of Teleogryllus commodus (Walk.) Harvard Univ. Press, Cambridge, Mass. (Orthoptera: Gryllidac). Austral. ]. Zool., 15: --. 1963b. Reply to criticism by R. D. Alex­ 541-545. ander. Syst. Zool., 12: 204.-206. 40 THE QUARTERLY REVIEW OF BIOLOGY [VOLUME43

McFARLANE,J. E. 1966. Studies on group effects SAEKI, H. 1966a. The effect of the population in crickets. III. Wing development of Gryllodes density on the occurrence of the macropterous sigillatus (Walk). Canad.]. Zool., 44: 1017-1021. form in a cricket, Scapsipedus aspersus Walker MEGALOV,A. A. 1924. [Observations on Gryllus (Orthoptera: Gryllidae). Jap. ]. Ecol., 16: 1-4. deserttts.] Trans. 4. All-Russ. Entmol.-Phyto­ [In Japanese, English abstract.] pathol. Meet., Moscow: 96-100. [In Russian.] --. 1966b. The effect of the day-length on the METCALF,C. L., w. P. FLINT, and R. L. METCALF. occurrence of the macropterous form in a I 951. Destructive and Useful Insects. Their cricket, Scapsipedus aspersus Walker (Orthop­ Habits and Control. McGraw-Hill, New York. tera: Gryllidae). Jap. ]. Ecol., 16: 49-52. [In MooRE, J. A. 1944. Geographic variation in Rana Japanese, English summary.] pipiens Schreiber of eastern North America. SELLIER,R. 1954. Recherches sur la morphogenese Bull. Amer. Mw. Nat. Hist., 82: 345-370. et le polymorphisme alaires chez !es Orthop­ teres Gryllides. Ann. Sci. Nat., llth Ser., 16: MORALES, A. E. 1940. Los Gryllotalpinae de 595-740. Espana. Bol. I'at. Veg. Ent. Agric. (Madrid), SoNAN,J. 1931. [Studies on Brachytrnpes porten­ 9: 212-233. tosus, Licht. and methods for its control.] Bull NowosIELSKI, J. W., and R. L. PATTON, 1963. Dept. Agric. Govt. Res. Inst., Formosa, 86: 1-22. Studies on circadian rhythm of the house [In Japanese.] cricket, Gryllus domesticus L. ]. Insect Phys­ THOMAS, E. and R. D. ALEXANDER.1957. iol., 9: 401-410. s., Nemobius meloditts, a new species of cricket --, and --. 1964. Daily fluctuation in the from Ohio (Orthoptera: Gryllidae). Ohio ]. blood sugar concentration of the house cricket, Sci., 57: l 48-152. Gryll11sdomesticus L. Science, 144: 180-181. --, and --. 1962. Systematic and behavioral NOWOSIELSKI,J. w., R. L. PA'fTON, and J. A. studies on the meadow grasshoppers of the NAF.GF.LE.1964. Daily rhythm of narcotic sen­ Orchelimum concinnum group (Orthoptera: sitivity in the house cricket (G. domesticus) and Tettigoniidae). Occas. Paper No. 626, Univ. the 2-spotted spider mite, Tetranychus urticae Mich. Mus. Zool. 31 p. Koch. ]. Cell. Comp. Physiol., 63: 393-398. THORNTHWAITE,C. W. 1933. The climates of the OHMACHI,F., and MASAKI. 1964. Interspecific s. earth. Geogr. Rev., 23: 433-440. crossing and development of hybrids between UMEYA,Y. 1950. Studies in embryonic hibernation the Japanese species of Teleogryllus (Orthop­ and diapause in insects. Proc. Imp. Acad. Japan, tera: Gryllidae). Evolution, 18: 405-416. 26: 1-9. OHMACHI,F., and MATSVURA.1951. [Four life s. VALDEYRON-FABRE,L. 1955. Observations snr la cycle types in Grylloidea.] Jap. ]. Af>Plied Zool., biologie de Brachytrnpes megacephalus Lef. en 16: 104-110. [In Japanese.] Tunisie. Rev. Path. Veget. et d'Entomol. Agric., PACE, A. 1967. Life history and behavior of a Fr., 34: 136. fungus beetle, Rolitothe1us cornutus. Occas. WALKER,T. J. 1962. The taxonomy and calling Paper No. 653, Univ. Mich. Mus. Zoo!., 15 p. songs of United States tree crickets (Orthop­ l'mn Y, M. 1832. Insecta Brasiliensia. Orthoptera. tera: Gryllidae: Oecanthinae). I. The genus Delectus Animalium Articulatorum Brasil, 116- Neoxabea and the niveus and varicornis groups 124. of the genus Oecanthus. Ann. Entomol. Soc. !'ICKFORD,R. 1953. A two-year life-cycle in grass­ Amer., 55: 303-322. hoppers (Orthoptera: Acrididae) overwinter­ --. 1963. The taxonomy and calling songs of ing as eggs and nymphs. Canad. Entomol. United States tree crickets (Orthoptera: Gryl­ 85: 9-14. lidae: Oecanthinae). II. The ntgr1cornis RAKSHPAL,R. 1962. Diapause in the eggs of group of the genus Oecanthus. Ann. Entomol. Gryllus pennsylva11ic11s Burmeister (Orthop­ Soc. Amer., 56: 772-789. tcra: Gryllidae). Canad. ]. Zool., 41: 179-194. --. 1964. Cryptic species among sound-produc­ --. 1964. Growth and respiration during post­ ing cnsiferan Orthoptera (Gryllidae and Tetti­ embryonic development of a diapause species, goniidac). Quart. Rev. Biol., 39: 345-353. Gryllus veletis (Alexander and Bigelow) (Or­ WEST, M. J., and R. D. ALEX'INDER.1963. Sub­ thoptera: Gryllidae). Indian J. Entcnnol., 26: social behavior in a burrowing cricket Anuro­ 67-75. g1yllus muticus (De Geer) (Orthoptera: Gryl­ RICHARDS,T. J. 1952. Nemobius sylvestris in S. lidae). Ohio ]. Sci., 63: 19-24. E. Devon. Entomologist, 85: 83-87, 108-111, WHEELER,W. M. 1893. A contribution to insect 136-141, 161-166, embryology. J. Morph., 8: 1-160. MARCH 1968] LIFE CYCLE ORIGINS AND SPECIATION IN CRICKETS 41

--. 1900. The habits of Myrmecophila nebras­ WYNNE-EDWARDS, V. C. 1962. Animal Dispersion censis Bruner. Psyche, 9: l ll-ll5. in Relation to Social Behavior. Oliver and WILSON, E. 0. 1963. Statement read before the Boyd, Edinburgh. U.S. Senate Subcommittee on Reorganization ZEUNER, F. E. 1939. Fossil Orthoptera Ensifera. and International Organizations, on October 321 p. Bull. British Mus. (Natur. Hist.), 8, 1963. London. A LIST OF CORRECTIONSA.ND EXPLANATORY-NOTES

for the paper:

Alexander, Richard D. 1968. Life cycle origins, speciation, and related phenomena in crickets. 4uart. Rev. Biol. 43(1):1-41.

p. 1. The number of "known" cricket species is optimistic, on the basis of published descriptions and recent arrangements of old names, partly because of undescribed species known to me and partly because my experience suggests that, accidentally or not, a higher proportion of the old names represent good species. It might have been better to date the Gryllidae to the Triassic Period. p. 3, table 1. I should have used the ·commonname "scaly crickets" for the Mogoplistinae. p. 6, col. 2, 1. 39. Ohmachi and Matsuura (1951) list cycles for several additional species in Japan. I am indebted to Sinzo Masaki for furnishing a translation of this almost unobtainable paper. p. 7, fig. 4 caption, 1. 6. Alabama, Georgia, and North Carolina ••• p. 10, col. 2, 1. 43-46. !_. commodus may indeed be non-diapausing in the north (I heard several males singing on 16 July 1968 at Rockhampton, Queensland). T. commodus and T. oceanicus, however, cannot be considered members of the same-species -- the latter simply a non­ diapausing northern race -- for both species were singing in the same field at Rockhampton without evidence of interbreeding. p. 21, fig. 10. The phrase "mostly interspecific" under "Hereditary variations" refers to most of the observed variations, and is not intended to imply that intraspecific hereditary variations in these regards are largely absent. p. 24, col. 1, 1. 39. A given area or habitat unit ••• p. 25, fig. 12. The phrase ''Most days and nights below 50°F." applies only to the period following September 27. p. 35, fi3. 15. In the figure itself "(a)" should apply to Scapsipedus aspersus and "{b)" to Gryllus fultoni, as can be discerned from the text. p. 37, col. 2, last line. Angus Hull (not Hall).