Mitt. Münch. Ent. Ges. Methods and Materials Insects and breeding conditions FromOctober1989toApril1990andfromNovember1991 toMarch1992threetrapswereputoutina10m X 20m area of a semihumid meadow poor in Vegetation (1430m o.s.l., Kilimanjaro area, Tanzania). The traps were checked three times a day at 7 am, 2 pm and 6 pm. Filter papers soaked with an acetone- cantharidin-solution were placed under gauze to prevent beetles from direct access. The baits were renewed every three to four days. Parallel to the trap controls, temperature and relative humidity were noted. The beetles were kept in 20 cm x 20 cm plastic boxes under field conditions in Kidia/Old Moshi (KiUmanjaro, 1430m o.s.l.). The two Fonn/fo»7//s-species were separated and bred on material (laterised red soll on volcanicrock, Downie 1972) taken from the habitatofthebeetles. Daily moistening with water providedsufficienthumidity. Adultbeetlesaswellasthelarvaewerefedwithfishfood (TetraMin)which consistsmainlyoffishmeal,pulverisedanimals,colouringandvitamins.Theuptakeofthefoodwaseasily Seen by the reddish droppings of the beetles and the gut contents of the translucent larvae. Notes on the behaviour resulted from individuals kept in the laboratory as well as from free living specimens. Vegetation releve Thephytosociological datawerecollected foUowing themethod ofBraun-Blanquet(1964). Inarepresent- ative, floristically homogeneous area of a few square metres in xerophilic meadows to some hundred Squaremetresinforeststhecoverabundancewasestimated,separated instrata,accordingtothefollowing Scale: r = Single specimens occurring in the reference area + = few species with small cover 1 = species with less than 5 % cover 2 = any number of species with cover 5-25 % 3 - any number of species with cover 25-50 % 4 = any number of species with cover 50-75 % 5 = any number of species with cover of more than 75 %. Feeding experiments with cantharidin All baits used to attract anthicids in Old Moshi (Kilimanjaro area, Tanzania) were put under gauze to prevent an direct access to cantharidin. At the end of the field studies about 50 specimens in April 1990 and 30 specimens in March 1992 ofboth Fonn/ro;/n(s-species werebrought alive to Germany. Because the beetleswerecapturedduringtheperiodofJanuaryand Februaryofbothyears,mainlyfemalesand hardly any males could be analysed. In the laboratory a part of the beetles had 9 days access to cantharidin. On the ninth day of cantharidin uptake the specimens were frozen (-55°C). Chemical analyses of cantharidin contents For each specimen an elytral fragment, a leg (middle leg with tarsus, tibia, femur and trochanter), one testicle, one accessory gland and one ovary were isolated. Dry weights of the fragments were registered (Sartorius semimicro balance R 200 D). The elytral fragment and the leg were directly used for the determinationofthecantharidin titres. Incaseofthepaired structures, testicle,accessory gland and ovary the fresh organs wereused foranalysis fortoo little material remained afterdrying. Only thevalueofthe dried paired structures was used for calculation. The beetle fragments were placed into injection needles (Pressure Mini-Injector 1 |al, Precision Sam- plingCorporation)andinjectedintoaCarloErbaVegaSeries2GC6000gasChromatographequippedwith 82 a CP-Sil 19 CB fused silica capillary column (Chrompack, 12,5m x 0,32mm I.D.; 0,25 m phase thickness; temperature program: 80°C to 280°C with a heating rate of 15°C/min; carrier gas: helium, 40KPa) connected to a Finnigan MAT Ion Trap Detector (ITD). Electron impact ionization (EI, 70eV) provides mass spectra with a characteristic fragmentation ofcantharidin: thebase peakwith m/z 96 and a fragment with m/z 128. Thecantharidin contentsoffield-caughtand cantharidin-fed specimensweredetermined by meansof calibrationcurves.Thecantharidincontentsofthehaemolymphresulted fromadditionand divisionofthe titres of the middle leg with the elytral fragment. With this mcthod only unbound cantharidin was analysed while bound cantharidin, which is also present in the structures, can not be detected by chemically untreated beetle tissue. Results Description and geographica! distribution of Formicomus rubricollis and F. gestroi Forwicomiifi rubricollis LaFerte (1848) shows a reddish pronotum with a dark head. Theelytra are black to greenorblue withan apparentmetallic lustre. As domost speciesofthegenusFormicomus, males possess teeth at the tibia and femur of the fore legs, their average length is 3,3mm. This species occurs in Central (Zaire), East (Kenya (Uhmann 1989), Tanzania (Uhmann 1981, Ueimann 1990), Ruanda, Burundi) and South Africa [Mozambique, Sambia, Simbabwe, Botswana, Namibia, South Africa (Uhmann 1984)]. Formicomus gestroi Pic (1894) is a very variable species. Colouring of the adults varies from brown to black, whereas the pronotum mostly is reddish and a vertical nut may be present. The elytra show often ametalliclustre.Themaleshaveteethatthetibiaandthefemuroftheforelegs.Theaveragelengthisabout 3 mm. The species occurs all overAfrica south of the Sahara. It is also found in East Africa (Kenya (Uhmann 1989), Tanzania (Uhmann 1981, Uhmann 1990), Burundi, Ruanda, Uganda), further north in Sudan and southward in Malawi, Angola,Sambia andSouthAfrica. EveninCentral (Zaire) and WestAfrica (Ghana, Gambia) Formicomus gestroi is widespread. Habitat Formicomus gestroi and Formico)nus rubricollis are sympatric in the area of Kidia/Old Moshi (Kilimanjaro area,1430mo.s.l.). Bothspeciesarefrequentinopenland. Densepopulationsarepresentonmeadows(see Vegetation releve) and at the rims of bushes with herb Vegetation. Aggregations of these beetles can be found on the herb Lepidiutn boiiariense L. (Brassicaceae), which reaches heights up to 0,5m. A Vegetation releve characterises the habitat of the two For;;;/ro;??)(S-species where most of the traps were put out. Releve of a semihumid meadow poor in Vegetation in Kidia/Old Moshi (Kilimanjaro 1430mo.s.l.). Cyperus rigidifolius Steudel Agerntutn coin/zoides Linne Hydrccotylc mawiii Htx3K.f. Apium Icptophyllum (Pers.) Benth. Oxalis latifolia H. B. & K. Bidois pilosa Linne Centella asiatica (L.) Ukb Caucalis incognita Norman Cyperus brevifolius (Rottb.) Hasskn. Com/za ßoribuiida H. B. K. Vigna parkeri Bak. Justiciaflava Vahl Zephyranthes graudiflora Herb. Justicia striata (Kl.) Bullock Dichondra repens ]. R. & G. Forst Phyllauthus suffrutescens Pa\ Eragrostis schwcinfiirthii Chiov. Spilauthes mauritiana (Pers.) DC. Oplismenus hirtellus (L.) P. Beauv. Tagetes mimita Linne Lepidium bonarieiise Linn6 Thunbergia alata Sims 83 Fig. 1: Lcpidiiun boiiariciisc L. parasitized by a mildew fungus. Diet Lcpidium bouaricnse is a frequent plant in the region. It often was parasitised by a whitish mildew fungus at the lower leaf surface (Fig. 1). The fungus produced orange sporangia. Investigationsofthegutcontentoffield-caughtspecimensofbothFormkomusspeciesrevealedthatthe whole intestines were füll of these orange sporangia. In the crop the sporangia could be found nearly in intact packs, whereas further down in the stomach and in the mid gut area garish fatty orange drops appeared. In the end gut the orange colour slowly disappeared and the gut content was soluble in water again. In thefield beetlescould beobserved tuggingwith their mandiblesat theleavesofLcpidium bonnrieuse parasitised by the fungus. Probably the beetles aggregate on these plants to take up the sporangia of the fungus. However, in the laboratory individuals of both Fonnicoiims species could also be bred with fish food (Tetra Min), maize porridge, bread, fruit and dead insects. The abdomen of dead Formicomus specimens within the plastic boxes were usually hollowed out. Mating behaviour On contact males of both species immediately jumped at any individual of the same species. Only after contactthemalesrealisewhetherthejumped onspecimenisa femaleornot. Ifthefemaleisreadytomate copulation takes placeatonce. Otherwise the femaletries to removethemaleby pushing itwith the hind legs. If the female is ready for mating the male hooks its fore legs into depressions at both sides of the prothoraxofthemate.Projectionsofthetibiaeofthemalesfitintotheseprothoracalnotchesofthefemales. Themaleprojectsitsaedeagusandinsertsitintotheupwardbendopened genitalvalvesofthefemale.The 84 femaleholdsitsantennaetLghtlypressedtothebodywhereasthemaleholdsthemhorizontallyawayfrom thebod\^ During copulation themale shakes thefemale 5e\eral times. Thepairremainsmounted atleast 24 hours. .Aiter se\-eral %veeks of capti\-it\- copulating pairs could be obsened again. Cantharidtn-attracted beetles tried intensi\'eh' to copiilate. The sp>ecimens jumped on each other, and often fourto six beetles got entangled. E\'en whilebeing collected trapped beetlesin small vials, indi\-id- ualsmounted each otherand copulated. Thisbeha\-iourwasnotobserx'ed soofteninindi\'idualsdirectl\' collected from the Vegetation. Reproduction Threetofoiirweeksaftercopulationthefemaleproducesforseveraldavsoval,whitisheggs.Theeggsare laid singlyorin small heapsintotheupperlayersofthe soil. Anisolatedfemaleproduced a maximumof 25 eggs per day. In choice tests wiih various plants (also Lq.ndium boiwriens-e, ^^•ith leaves, flowers, stems and roots) and soilniaterial Comingfrom thehabitatofthespecies,eggscould alAvavsbefound inhollow mm Spacesintheupperlayerofthesoil.After12daysthewhitish,Httlesclerotisedlar\-ae,of1 lengthand 0^mm \Nidth, hatch. Thelar\"ae,withabdomenfirst,lea\"etheeggskinh\\\Tndingandbeatingmovements.InthefoIlo"v\Tng 6-8weekstheno^^•%vhiteanddistinctlysclerotisedlar\'aegroivtoama.ximumlengthof6mmandawdth mm of0,6-1 (seetable1). Lar\'aeform a longish earthen puparium shortlvbeforepupation. The develop- ment from the sluggish lar\'ae forming the pupariimi to the pupa exarata takes eight days. Afterthepuparium isfinished thelar\-aebecomesstiffexceptforbeatingmovementsoftheabdomerL Three days laterpupation can benoted. After some days theAvhite pupa tums vello's%ish. On mechanical Stimulation the pupa reacts \\-\\h beating mo\ements. After 10 days the adults emerges. Freshlv hatched beetles are immediateh' canthariphilous. Table1. EVevelopmentofthelar\-aeofFormicomusgestroiandFormicomusrubricolUs.Days:Ageofthelar\'aeafter hatching; Length: Average length of the larvae, means vnih Standard deviation; VVidth: Maximum wdth of thorax, means vnih Standard de\'iation; n: Number oflarx^ae; -: No values. Davs 10 days 3-4 weeks 6-8 weeks 12 days Fig.2: LifecycleofFormicomusgestroi. Lifespanoftheadultsin thelaboratory: max. 5 months;ovipositionafter copulation:about3-4weeks; hatchingofthelarvaeafteroviposition:12days; developmentofthelarvaetopupa: about 6-8 weeks; pupa: 10 days. Table2. Choice fest with larvae of Formicomus: Number of larvae in the different soll types. Days: Age of the larvae after hatching; n: Total number of found larvae; initial number oflarvae: n=30 Days Novembern 1Dece~\mberr 1 FIebruarry Fig.3: Temperatureandrel.humidityinKidia/OIdMoshi(Kilimanjaro1430mo.s.l.,Tanzania)at7hr, 14hrand 18 hr in the period ofNovember to February 1991/92. In Upper layers of fertilised soil from cultivated areas, that had been passed through a sieve, Fonni- comus-larvae ofall sizes were found. Probably the larvae develop quickly specially in loose soil fertilised with organic material. Similar conditions could be noticed also for the species Formicomiis pedestris. This beetlespecieswastrappedinwheatfieldsintheMediterraneanregionwhichwasfertilisedwithstalllitter. Breeding of this Foruücomus-species probably was not successful because the larvae were not fed with a suitabledietin thelaboratory(Schutz 1989). Fishfood giventotheAfricanFormicomus-larvaein following breeding experiments obviously was the adequate diet. Thus larvae of F. gestroi and rubricollis grew rapidly, the larvae pupated, and the adults hatched. Phenology During the field studies in Kidia/Old Moshi (Kilimanjaro area, Tanzania) in the months ofOctober 1989 to April 1990 and November 1991 to March 1992 temperature and the relative humidity were noted daily at 7 am, 2 pm and 6 pm (Fig. 3). About the same time the cantharidin traps were checked. In captivity the beetles laid eggs mid-November, end of December, mid-January and beginning of February.TheovipositioninDecembercametogetherwiththesmallrainyseason,thoseofNovemberand January with the dry seasons. Notably the months ofJanuary and February are the hottest and the driest periods of the year (see Fig.3). From November to February specimens of f. gestwi and F. rubricollis were continuously attracted to cantharidin traps(seeFig.4-7). Amaximumwasachieved duringtheperiod fromNovembertoDecember. Few beetles were trapped end of December/beginning of January 1989/90 as well as 1991/92. Heavy rainfalls during the small rainy season probably caused a decrease of the population density (see Fig. 3). Temperatures feil to 15°C in the morning hours and the relative humidity increaseci to over 80 %. Even during the midday hours (2pm) the temperatures only reached about 20-22°C with a relative humidity that did not feil below 80 %. Similarweatherconditions, that means comparatively low temperaturesata high relative humidity, also predominated during the small rainy season in the evening hours. 87 November December January February March November December January February March Figs4-5; AbundanceofFonnicoiniisgcstroi and nibricoUis 1991/92 fromNovember toMarch in Kidia (Kilimanja- ro-area, Tanzania). ?: No values; 0: No catches in cantharidin traps. Table3. Weekly sums and sex ratio of cantharidin attracted Formicomus gestroi and F. rubricoUis. Fgl: Weekly sums Formicomus gestroi 1989/90; Frl: Weekly sums Formicomus rubricoUis 1989/90; Fg2: Weekly sums Formi- comusgestroi 1991/92; Fr2: WeeklysumsForm/comws rwbnco///s 1991/92; m%, f%: Proportionofmales, females of the weekly sum; -; No values. Week 10 8- Formicomusgestroi1989/90 T3 6 4 - i - 2 n O-L ? ? ? ? ? ? 7 ? November December January February March April 10 8 - Formicomusrubricollis 1989/90 6 - - 2 0? November December January February March April Figs6-7: AbundanccofFonnicotniisgL'stroiand ruhricoUis 1989/90from NovembertoApril in Kidia (Kilimanjaro- area, Tanzania). ?: No values; 0: No catches in cantharidin traps. An increase of trapped beetles was noted again at the beginning of the dry season in January, but the I numberofcaught specimens was notsohighanymore duringNovemberand December. The mean daily 1 temperatures were in average lower than during the rainy season (15,9°C during the week from the 5.1. ', tothe11.1.92compared to16,4°Cduringtheweekfromthe22.12.tothe28.12.91)andtherelativehumidity ' feil for about 10 %. Over midday temperatures raised to 30°C and the relative humidity sank to 49,7% I (5.1.-11.1.92)and remained relativelylowuntil theeveninghours (58,7% duringtheweekofthe5.1. tothe 11-1.92-74,3 % during the week of the 22.12. to the 28.12.91). From the beginning of February few beetles j stillwereattracted.Frommid-Januarytheportionoftrappedfemalesnoticeablyincreased.ForF. rubricollis I specimens the proportion of females was generally higher, but also in this species from end of January 1 onward the number of females attracted raised from 66,7% (26.01-1.2.1990) to 90% during the week of ( 9.2.-15.2.1990. From mid-February to the beginning of March no beetles at all were trapped. From the I middleofMarchand inAprilwiththebeginningofthegreatrainyseasoninboth fieldstudyperiods(1990 I ' and 1992) beetles appeared again (Fig.4-7). Analyses of cantharidin contents ThecantharidincontentsofvariousstructuresofthebothFor//;;'a)/H;<s-speciestrapped in Kidia/Old Moshi were determined by GC-MS (gas chromatography-mass spectrometry), because first analyses had shown that the structures mostly contained only small amounts of cantharidin. Therefore it was offen necessary to identify cantharidin with the help of the diagnostic mass spectra fragments. On the one hand beetles with no access to cantharidin in the laboratory were analysed, on the other individuals that had had the possibility to take up synthetic cantharidin for 9 days in the laboratory. J Whilepreparing malesitwas found conspicuously, thatthepaired testesshowed enormousdifferenc- I es in size. When a spermatophore was present in the abdomen ofthe male, one ofthe testicles was small i (abouthalfofthesizeoftheother)and ofamorphousshape.Theothertesticleshowed thenormal3-armed shape. Freshly hatched males where a spermatophore was not yet formed in the abdomen always \ possessed equally shaped 3-armed testes. j 89