EGG HATCHING AND LIFE HISTORY OF A FAIRY SHRIMP BRANCHINECTA MACKINI DEXTER (CRUSTACEA: ) IN A MOHAVE DOCKET PLAYA DRY LAKE)' DESERT (RABBIT 11-AFC-2 LARS CARPELAN LON R. BROWN AND H. DATE Universityof California, Riverside, California 92502 RECD. MAY 01 2012 Abstract. Observations on populations of Branchinecta mackini in temporaryponds, which last for as little as 3 days to as long as 4 months, on a desert playa showed that hatching of dehydratedcysts (eggs) follows initial entry of water into the basin if salinity remains low. Salinity ranges from an initial 0.5%o to as much as 34%o as the pond evaporates. Hatching is continuous at constant low salinity, but since salinity generally increases rapidly, initial hatch is usually of short duration. Additional periods of hatching follow further inflows of water or after melting of ice, that is, after reductions in salinity. The duration of hatching is in- versely proportional to rate of increase in salinity. When salinity of small-volume summer ponds increases at rates above 500 ppm (1,000 Vmhos) per day, there is virtual inhibitionof hatching. Laboratory studies showed that egg hatching was controlled by both salinity and oxygen operating in various combinations to inhibit or stimulate hatch. The hatching characteristics of desiccated eggs collected from the dried mud of the lake basin differedfrom non-dried eggs obtained from laboratory cultures after ejection from ovisacs of living females. The findingthat a salinity-oxygencomplex regulates hatching in a desert pond permitstenta- tive explanation of a differencebetween branchiopods of humid and arid regions. In both cases the branchiopods are characteristicof astatic waters, and stimulation of the egg to hatch must be by some factor that changes at the time of origin of the temporary pond. In humid regions the factors of concern are temperatureand oxygen. If the water undergoes signifi- cant changes in salinity,as it does in arid regions, control of egg hatch may be by both salinity and oxygen, with temperaturelimited to control of rate of development.

INTRODUCTION salinityaffects the rate of excystmentof Artemia for The Anostraca (fairyshrimps) are branchiopod salinc, a specialized Anostracan,exceptional waters. The con- whose usual habitats according to its adaptationsto hypersaline fromits dormantstage by ex- Dexter (1959) are "rain pools and temporary trol of emergence has been clarifiedby Clegg pondsthat form from melting snow and ice." The ternalosmotic pressure there are reportsof osmotic dormantstage, which is usually an embryonated (1964). Although ionic adaptations of active stages of other egg or cyst,permits survival during periods when and Cole and Brown the ponds are dry. Because mere wettingoften branchiopods(e.g., Broch 1969, seems to have writtenthe is not sufficientto bringabout resumptionof de- 1967), Horne (1967) only paper reportingthat salinity inhibits the velopment,various factorsor combinationsof fac- offreshwater Anostraca in naturalwaters. tors, reviewed by Hutchinson (1967, p. 562), hatching of freshwaterAnostraca have been have been suggestedas stimulifor hatching. Fail- Most studies made not in arid but in humid regions, where ure to develop has sometimesbeen attributedto changes in salinityare negligibleand where reg- an obligatediapause, but it is morelikely an adap- ulationof hatch is predominantlyby oxygencon- to stimuliother than wetting,thereby in- tation centration. Thus Broch (1965) foundthat stim- under certainconditions un- hibitingdevelopment ulation of the dormantstage of Chirocephalopsis completionof life cycle. favorableto bundyi(Forbes) is by low oxygenconcentration. below was foundthat In the study reported it Low oxygentension can also inhibitexcystment as eggs of Brachinectamackini (Dexter) hatchingof reportedby Moore (1963), who foundthat eggs desertpool was regulatedmainly by in a Mohave of two speciesof Anostraca-Streptocephalusseali of oxygen changes in salinity,with involvement Ryder and holmani (Ryder) influenceof temperature. tensions, and without would hatch only in well-aeratedwater. Bishop well withenviron- The lifecycle is so synchronized (1967) foundthat low oxygen concentrationsin- mentthat thereis littlewastage of eggs. hibit hatching of the conchostracan,Limnadic that Jenningsand Whitaker (1941) reported stanleyanaKing. Brewer (1964) reportedthat 1 Supportedby National Science FoundationGrants lowering of oxygen tension is the stimulusfor on the Ph.D. GB 3500and GB 6668. This paperis based hatchingof a calanoid copepod, Diaptomus stag- thesisof L. R. Brown. Manuscriptreceived December 18, 1969; acceptedAugust 8, 1970. nalis Forbes. Sussman and Halvorson (1966, 42 LON R. BROWN AND LARS H. CARPELAN Ecology,Volume 52, No. 1 p. 201) give even greatergenerality to oxygen as FIELD STUDIES ab- a stimulusin theirdiscussion of presenceand Introduction sence of oxygenas an environmentalinfluence on germinationof spores. This reportis restrictedto ponds appearingin Need for combinationsof factorshas been re- one basin on one dry lake bed on the Mohave ported. It has been known since 1941 that low Desert. Rabbit Dry Lake is in the Lucerne Val- oxygen stimulateshatching of aedine mosquito ley at an altitudeof 2,950 ft (950 m) in the south- eggs (Gjullin, Hegarty,and Bollen 1941). But easterncorner of the Mohave Desert in San Ber- Horsfall (1956 & 1958) showed that additional nardinoCounty, California (34?27' N, 117? W). factors,such as time for developmentand proper Althoughother California desert ponds have been temperature,are neededto "condition"Aedes eggs studied,this pond is chosen for reportbecause it beforethey can be stimulatedto hatchby low ox- was the most intensivelystudied (data are avail- ygenconcentrations. Interactions of two or more able fromsix differentfillings, during three sea- factorsare thus possible, and Weissman-Strum sons over a two-yearperiod) and also because it and Kindler (1963) have postulateda two-stage is the simplest biologically. One species of process of hatchingof A edes eggs that requires branchiopod-B. mackini Dexter-predominates low oxygen to stimulatebreakdown of a water- almost to the exclusion of other macroscopican- permeablebarrier, which then permitsthe entry imals. B. mackiniwas firstreported from brack- of water,an osmoticphenomenon. ish lakes in the state of Washingtonby Dexter It thus seems thatapparently conflicting reports (1956) and has since been reportedin California of what stimulatesand what inhibitsarthropod and Nevada (Dexter 1959). excystmentare due to specificdifferences, due to In nearby Mohave waters there are other involvementof more than one factor,or due to branchiopods,including another Branchinecta sp. all eggs of one speciesnot beingin the same "con- (B. lindahli) as well as Thamnocephalusplatyurus dition." The differencesin stimuli required by and Triops longicaudatus(which are oftenasso- humidregion (freshwater)anostraca such as Chi- ciated), togetherwith complex communitiesof rocephalopsis,which respondsto low oxygenand Cladocera, Ostracoda, Conchostraca,and Insecta. temperature,and the arid region (hypersaline) In Rabbit Dry Lake, the only macroscopicin- Artemia, which responds largely to osmotic vertebratein additionto B. mackiniis Branchi- changes,suggested that studyof a species inhabit- necta gigas Lynch. Unlike mackiniwhich is pres- ing waters of intermediatesalinity might con- ent in all ponds formed-winter, spring and tribute to understandingthe requirementsfor summer-B. gigas is presentin winterand spring resumptionof developmentand the broaderprob- only. It apparentlyis intolerantof summerheat. lem of adaptationto environment. B. gigas was firstreported from Washington in To put the Rabbit Dry Lake waters into per- 1937 and since then fromMontana, Nevada, and spective,we can compare them with conditions Utah (Dexter 1959); it has not previouslybeen reported elsewhere. In the tropical desert of reportedfrom California. Despite its large size northernSudan, Rzoska (1961) reportedconduc- (50-90 mm) it may be missed in samplingbe- tance at the end of the existenceof a temporary cause of its relativelysmall numbers. It is a pred- it in a ratio of pond to be 1,130 ,mhos (about 550 ppm). This ator on mackini,which outnumber gigas to 40,000 mackiniin some collections. terminalconcentration is about equal to the min- one was observedto consume about 35 young imum initial salinityof the Mohave desert pond, Gigas mackiniper day. Prey smallerthan 1.5 cm are not much than salinities by and greater reported eaten whole, larger ones are nipped in two and Prophet (1963) in studies in Kansas and Okla- onlythe abdomeneaten. Both prey and predator homa, where no pond had a conductivitygreater hatch at the start of the pond, so predationon than 555 Vmhos (about 275 ppm). The waters mackinioccurs from the verybeginning. B. gigas of Rabbit Dry Lake increase in salinity from is mentionedhere only incidentally, but it obviously about 0.5%yoto terminalsalinities about equal to is not unimportantto the general ecology of the sea water (34%yo)during the finalstages of sum- pond and to the populationdynamics of B. mac- mer ponds. Horne (1967) foundthat the salinity kini. Gigas is ignoredin the followingaccount in in several Wyoming ponds increased to a max- orderto focusattention on the egg-hatchingof B. imumof from0.5 to 1o%, except for one pond in mackim. which Branchinectalindahli and Streptocephalus texanus occurredin which the salinityincreased Field methods to 16.5%o The latterconcentration is comparable Instruments.-Changesin temperature,specific to Rabbit Dry Lake. conductance,oxygen concentration,and pH were Winter 1971 EGG HATCHING OF A DESERT FAIRY SHRIMP 43 measured with the followingbattery-powered in- extend into September. Summer rains, mostly struments:Temperature: Yellow Springs Instru- in Julyand August,usually amount to a relatively mentCompany, Tele-Thermometer, Model 43TD; small fraction(1/10 to 1/4) of the annual total. Specificconductance: Industrial Instruments, Solu Standard air temperaturesin Julyand August Bridge,RB3-338, using conductivitycells G-2 and average 280C, with maximumof 410 and min- G-20; Oxygen concentration:Yellow Springs In- imumof 120C. In winterair temperaturesrange strumentCompany, Oxygen Meter, Model 51; froma minimumof -80C to a maximumof 200 pH: Beckman Instruments,portable pH meter, foran averageof 90C. Temperaturesbelow freez- Model N; Radiation: InstrumentSpecialities Com- ing can occur fromNovember through April on pany,Radiometer, ISCO Model SR. this high desert. Sampling.-Sampling of the shrimppopulation was accomplishedfrom shore either with a tri- General environment.- angular net (80-mesh nylon), mountedat right a. Morphometry.-RabbitDry Lake is a playa angles on the end of a 10-ftaluminum pole, or by (a "flat-flooredbottom of an undrained desert suction. The shrimpwere filteredinto a small basin") about 3 km long and 1 km wide lyingin 80-mesh nylon bag placed over the inlet of the a valley betweenthe Graniteand San Bernardino samplingvessel. The vacuum methodwas used Mountains. The lake bed is bisectedby a high- forsampling ponds whose depthswere 1-2 cm or way runningeast/west, which dams offthe south- less. The sampleswere preservedin 4% formal- ern quarterof the playa. The studypond is the dehyde for measuringand counting. The sizes deepestpart in this southend of the lake. Water reportedare lengthsof preservedindividuals from flowsinto it fromthe northernslopes of the San anteriorend to anus as measuredunder a stereo- BernardinoMountains, via ravines and canyons, scopic microscope with a calibrated eyepiece- but the valley is essentiallyarheic, and littlerain micrometer. falls on the valley itself. Estimation of date of hatching.-Because the It is seldomthat the entireplaya is coveredwith ponds were not under daily observation,the date water. In winterthe studypond usually is about of hatchingwas not always knownfor the shrimp 120 m long by 24 m wide, witha maximumdepth that hatched subsequent to the initial hatching of 20 cm, whichis not at the centerbut near the period. An estimateof subsequenthatches was northernedge of the long dimension,which paral- made fromthe growth rate of the initial hatch, lels the highway. A volume equal to that of the using the largestimmature individual collected on winterpond is seldom attainedin summer. Due a givenday, taking 10 mmas the usual size at first to the shallowprofile of the basin, a large fraction maturity. Growthrates of older matureindivid- (up to 34) of the area coveredby a large-volume uals (longer than 10 mm) could not be used, be- winterpond may be covered by a small-volume cause growthalmost ceases duringmaturation of summerpond, but with a very shallow sheet of the gonads. After maturation,growth resumes water,most of which is only 1-2 cm deep. Be- and then slows as maximumsize is reached,ac- cause of the extremeshallowness, the ponds do cordingto studiesof Dexter and Ferguson (1943), not last long; the surface-to-volumeratio favors Coopey (1950), and Moore (1963). The max- rapid evaporation. imum size of three-month-oldB. mackini adults b. Light.-The pond water is opaque with sus- does not greatly exceed 30 mm. Collections pendedsoil particlesso finethey do not settleeven showedthat hatching in RabbitDry Lake is seldom during windless periods or when ice covers the continuous. Though there may be a continuous pond. The colloidal-sizedparticles do precipitate size distributionin the population,frequency peaks near the end of a pond's existencewhen the sa- indicateseparate bursts of hatching. linityincreases rapidly (see below). Data on lightpenetration taken on a clear wind- Results less day between1200 and 1225 on April 23, 1967, Climate.-The studypond is in an arid region. are representativeof data gatheredat other sea- According to U. S. Weather Bureau records, sons. Columns 1, 2, and 3 of Table 1 show rainfallin Lucerne Valley during calendar years the energyfalling on the pond's surfaceand pen- 1965, 1966, and 1967 totaled 18.2, 6.1, and 11.1 etratingto depths of 4 and 10 mm. The 4-mm cm, respectively,for an average of 11.8 cm (4.65 depthwas selectedbecause the remotesensor used in.). Most rain fallsin Novemberand December. withthe radiometerwas just visibleat this depth. Rains in spring can be negligible,but rains in Columns4 and 5 show the energyat 4 and 10 mm April 1967 amountedto a thirdof thatyear's total. relativeto that at the surface. The amount re- May and June are generallyrainless as are Sep- flectedand re-radiatedwas not determined. It is temberand October, but the summer rains can obvious fromthese data that longerwave lengths 44 LON R. BROWN AND LARS H. CARPELAN Pclno-v vollutmne 52 Mn 1

TABLE 1. Energy transmittedto 4 and 10 mm depths, 25 Spring Pond, Rabbit Dry Lake, April 23, 1967

(1) (2) (3) (4) (5) 20 Energy-p.w/cm2per my. E Percent of energy '3. ______- at pond surface AtAtpond_ pond Uz 15 / my surface At 4 mm At 10 mm (2)/(1) (3)/(1) 380 29.1 0.110 0.036 0.38 0.12 400 50.5 0.000 0.002 0.00 0.00 105 425 58.9 0.000 0.000 0.00 0.00 450 81.6 0.000 0.000 0.00 0.00 475 90.0 0.005 0.001 <0.01 <0.01

500 86.1 0.038 0.063 0.04 0.07 0 'Oa 0 50 100 150 200 250 300 350 400 SMolilV(mOs 525 83.9 0.160 0.230 0.19 0.27 ! | ' @ '' I X. S e BI I I * Salinity (%.) 550 72.9 0.530 0.660 0.73 0.91 2.0 4.0 6.0 8.0 10.0 12.0 (NaCI) 575 76.2 1.600 0.730 2.10 0.96 600 73.3 4.700 2.300 6.41 3.12 FIG. 1. Relationof specificconductance to osmolality 626 66.6 5.300 2.800 7.96 4.21 and salinity,based on standardNaCl solutions(o) and 650 66.5 5.300 4.000 8.00 6.02 samplesfrom the SpringPond, 1967 (-). 675 62.0 11.200 4.900 18.20 7.91 700 51.1 10.900 5.700 21.30 11.10 725 45.6 7.800 6.000 17.10 13.12 Because all the ponds discussed are withinthe 750 40.9 8.200 3.700 20.00 9.05 same basin, the waters have the same ionic com- position. Notationsof salinity(%o) shown along the curves of Figures 3-6 are expressedas NaCl, make up most of the energypenetrating into the based on conductivitymeasured in the fieldcon- water. At no wave length does light penetrate verted to salinityby the graph in Figure 1. In below 20-25 mm. Table 2 are shown the ionic constituentsof the One effectof this turbidityis that the bottom waterin Rabbit Dry Lake in the range of salinity is dark, and hence photosynthesisis restrictedto of concernto the egg-hatchingobservations. It the uppermostlayer of wateror to the substratum is a chloridewater but withfairly high proportion at the shallow edges of the pond. Anothereffect of bicarbonate,which accounts for deviation of is that it is impossibleto observe the life in the in pond except as organismssurface or become vis- its curve from that of the chloride standard ible in the uppermost4 mm. Figure 1. c. Salinity.-Estimates of salinityare based on d. Dissolved Oxygen.-Oxygen was usually determinationsof conductivity.A graph for the near saturationat all seasons and at all times of interconversionof specific conductance in mi- day. In the shallow waters concentrationswere cromhos(,umhos) of the pond waterto osmolality the same at the surfaceand just above the bottom in milliosmols(mOs) and to salinityin parts per especiallyunder windy conditions. Occasionally, thousand (%oo) is presentedin Figure 1. The on windlessdays, supersaturation was observedin osmolalityof seven samples of water fromRabbit restrictedareas when the planktonicalgae, which Dry Lake (Spring Pond, 1967) and of a standard are predominatelyflagellated single cells, rose to NaCl solution were determinedwith a line- the surface where they formed visible green operated freezing point osmometer (Advanced streaks. Within such a streakthe oxygen could InstrumentsCo., Series 63). be above saturation(e.g., 112%), while just out-

TABLE 2. Chemical analyses of three samples of Rabbit Dry Lake water, at 1.4 %O (2,200 tmhos), at 3.5 %O (5,600 Vimhos),and at 6.0 %Y (10,000 Vmhos)

Analysisat 1.4 /o/ | Analysisat 3.5 o/0o Analysisat 6.0 ?/oo

O/0oratio ?/noratio ?O/oratio Ion Meq/1 ?/oo ion/total Meq/1 ?/oo ion/total Meq/1 /oo ion/total

Na ...... 22.0 0.506 0.374 52.5 1.210 0.347 95.0 2.200 0.369 K ...... 0.15 0.006 0.004 0.14 0.005 0.001 0.26 0.010 0.002 Ca ...... 0.38 0.008 0.006 0.65 0.010 0.003 1.10 0.020 0.003 Mg ...... 0.02 0.002 0.0001 0.14 0.002 0.0006 0.31 0.004 0.0007 C1 ...... 14.60 0.520 0.384 38.69 1.380 0.396 68.03 2.400 0.403 S04 ...... 1.30 0.062 0.046 3.33 0.160 0.046 6.04 0.290 0.048 CO3 ...... 0.8 0.050 O.C37 2.96 0.170 0.049 5.36 0.320 0.054 HCO3...... 4.50 0.202 0.149 12.30 0.550 0.158 15.90 0.710 0.119 Winter 1971 EGG HATCHING OF A DESERT FAIRY SHRIMP 45 side and below the green area the concentration were 30 and 220, withdiurnal fluctuations as great would be at saturationagainst air. as 170 and as small as 70. During windlessperiods and when salinitywas Figure 2 indicatesthat temperaturesin spring greaterthan 10%o,the algae settledout together were generallyintermediate between winterand withthe suspendedsediments. Bubbles of oxygen summervalues. Smallestranges occurred in win- then appeared on the bottomin localized areas of ter, with more variable conditionsin spring and algal concentration,and the clear overlyingwater summer. B. mackini,present in all these condi- showedoxygen supersaturation up to 166%. Ox- tions, is obviouslyeurythermal, swimming under ygenconcentrations as low as 65% were observed ice in winter and also active at temperatures under ice. Still lower values occurred at the greaterthan 300 in summer; it was observedto mud/water interfaceand in the bottom mud. withstanddaily changesas greatas 190. Conditionsat the bottomwill be discussed in the sectionbelow on the microenvironmentof the egg. Microenvironmentof the egg.- e. Hydrogenion concentration(pH).-The pH a. Dried state.-When the pond basin is drythe was usually 8.8 to 9.2 duringthe entireexistence eggs (cysts) are in the dried mud, not layeredat of a pond. Often there was no change greater the surfacebut distributedthrough the top 5-10 than 0.1 pH unit during any given day. Only mm. Two layerscan be differentiated:an upper- within localized areas of intense photosyntheticmost darker brown layer (A), a lighterbrown activity,as along the edge of the pond in a mat of layer (B), and the soil beneath(C). The surface algae or withina streakof suspendedalgae, were shrinkson drying,showing polygonal desiccation higherpH values observed. Above activelyphoto- cracks that extend down into C. The eggs are synthesizingalgae the pH at timeswas 10.2, while predominantlyin layers A and B, which remain values an entirepH unitlower (in the usual range togetherand separate from C when completely from8.8 to 9.2) were present 1-2 cm fromthe dry. The thicknessof combinedlayers A and B algal concentration. varies from5 to 10 mm. On additionof water, Existence of such localized conditionswas de- layers A and B swell to a jelly-likeconsistency; pendentupon lack of winds. Stirringby winds it is from these layers that the suspended sed- rapidlyreduced extremes of pH to the usual value iments (from layer A) and the slightlydenser near 9.0. At the mud/waterinterface pH was slurry (from layer B) are derived. The eggs 8.6, whilethe pH of the mud itselfwas 8.3. withinthese layers are fromtwo sources: those f. Temperature.-During the summer (August ejected duringthe life of the pond by living fe- 11 to 29, 1965) the minimumtemperature of the males, and those remainingin the brood pouches water was 120 and the maximum320 (maximum of femalesdying at the end of a pond. 390 recorded in summer 1966). The diurnal b. Changes at pond origin.-The pond can fill rangevaried from 120 to 190. In the winterpond rapidlywith runofffrom cloudburst-type rains in (November 25, 1965 to March 15, 1966) the the area. On one occasion it was observedto fill minimumtemperature (under ice) was 10 and the withinan hour by inflowfrom the southand east. maximumwas 14'C. The smallestdiurnal change Many eggs are initiallyin the topmostportion of the gelatinousslurry as the soil surroundingthem in winterwas 30 and the greatestwas 120. The goes into suspension. Currentscan carrythe eggs minimumand maximum temperaturesrecorded about untilthe pond is established. Then the eggs for the springpond (April 11 to May 9, 1967) settleinto the bottommud. The eggs do not float, eitherin the desiccatedor in the hydratedcondi- 40 0 Summer Pond 1965 tion. No eggs can be collectedby net aftera pond 6 Spring Pond 1967 is establishedunless the net dips into the mud. a Winter Pond 1965-66 Hence, the eggs must be stimulatedto hatch or on inhibitedfrom hatching by bottom conditions, whichmay be differentfrom the open waterabove, m 20 and are not homogeneouswith respectto oxygen r20 concentrationsnor, initiallyat least, to salinity. 4 15 As water flowsinto the dry basin,the eggs be- gin to absorb water. The desiccated eggs are deeplyindented on one side and become spherical only after taking up water, a process requiring 8000 0600 1200 1800 2400 TIME IN HOURS time. Laboratory observationsshow that at a FIG. 2. Ranges of diurnalwater temperatures in sum- constantsalinity of 0.5%o (1,000 pmhos), the first mer,winter, and spring. egg roundsup at the end of 50 min; at 10,000and 46 LON R. BROWN AND LARS H. CARPELAN Ecology, Volume 52, No. 1

20,000ptmhos it takes60 and 120 min,respectively. 5 - , 55.000 Tuhos 5 - '.O 55.000 ..mho. + presence of noupili The timefor half the sample of eggs (50 of 100) 2.5 %. to swell (at room temperatureof was 0 minimumsize 0.6 mm 20-25?C) 2 4 - * minimumsize 1.0 mm 88, 117, and 172 min at the threeconcentrations * minimumsize 2.0 mm .5 mentioned. p %. c. Salinity.-The salinity at the egg surface w -s l 1.4%. while in the bottommud is difficultto determine. 0 It has been assumed, because of their proximity 'C t1.1%. to the soil/waterinterface, that the eggs are sur- roundedby the same salinityas in the open water o Oiut ~ .4 , I I of the overlyingpond afterequilibrium has been attained. However, as in the case of the oxygen concentrationsdiscussed below, a salinitygradient C0 10l 20 30 40 50 60 70 80 90 may exist from the interfacedownward. The TIME IN DAYS most deeply-lyingeggs may thereforebe at a sa- FIG. 3. Changes in salinity and presence of young, linitygreater than that determinedin the water Summer 1965 pond. above. The existence of these gradients,which resultsin the eggs being in a heterogeneousen- 3) .-The slower increase in salinity from the vironment,helps accountfor all eggs not hatching secondto fourthdays was due to rain on the sec- at once,though variability of eggs is also involved. ond and thirddays. Afterthe rains salinityin- d. Oxygen.-It proved difficultto obtain satis- creasedrapidly, reaching 2.5%Oo on the ninthday. factorymeasurements in the microenvironmentOn the final (18th) day salinitywas about 34%0 surroundingthe eggs. Because of the sediments as indicatedby a specificconductance of 55,000 and the difficultyof obtainingsamples fromdefi- t~mhosin the remaining5 mm filmof water. nitedepths below the interface, the Winklermethod From maximumsizes collectedon the second, foroxygen could not be used. Betterresults were third,and fourthdays (2.5, 4, and 6 mm) the obtainedwith a polarographicoxygen probe (Yel- growthrate of earlystages is 1.25 mm/day. This low Springs InstrumentCo.), but because it con- rate is minimal,for by the tenthday 17% of the sumes oxygen,it must be kept movingto obtain populationwas 19 mm long, and the longestin- a steady reading. The probe can be positioned dividual was 21 mm, indicatingthat 2 mm per at desireddepths, but the movementrequired dur- day is about the maximumrate in summer. ing measurementmay stir in better-aeratedwater Nauplii were present up to the second day, fromabove and mud of loweroxygen content from absenton the third,reappeared on the fourth,and below. youngimmature stages were presentuntil the 11th A sharpreduction in oxygenwas foundjust be- day (Fig. 3). Those longerthan 10 mm on the neath the soil/waterinterface. While the open tenth day (80% of the population) probably waterof the pond had oxygenat saturationagainst hatched during the first two days (the initial air (e.g., 8 mg 02/1), the concentrationat the in- hatch); a second hatchaccounts for those shorter terface(with the membraneof the oxygen probe than 10 mm. The reappearanceof nauplii coin- restingon the softmud surface) was 4-5 mg/1. cided with dilutionand a slowed rate of salinity In the uppermostlayer of mud (the top 3-5 mm) increase. the oxygen concentrationwas 2-3 mg/1; in the The hatchingin the summerpond of 1965 oc- layer 5-10 mm below the surfacethere was less curred in two bursts, each followinginflow of than 1 mg/l. This findingis in keepingwith the water and endingduring rapid increasein salinity expectationaccording to the descriptionof Boul- in a shallow pond. din (1968) of an aerobic layer of mud above an WinterPond, November25, 1965 to Mtarch15, anaerobic layer. 1966. (Fig. 4) .-This pond lasted 111 days,which was the longest durationof any pond observed. O bservationson individualponds The initial salinityof 1,000 l~mhos (the dotted Each pond thatoccurred in the studybasin from line of Fig. 4) was estimatedfroni the runoffthat the summerof 1965 to the summerof 1967 is was fillingother ponds. There was reductionin describedbelow, but with emphasison one pond salinity when more water entered on the 14th for each season. Summariesof the findingsper- day; then a series of lightrains kept salinitylow tinentto egg hatchingare presentedin Figures until the 44th day. On the last day the remain- 3-6. Two short-durationsummer ponds are dis- ing film of water had a conductanceof 30,000 cussed,but findingsare not figured. iWmhos(about 18%o). Summer Pond, August 11 to 29, 1965. (Fig. On the 15thday (the firstsampling of the pop- Winter 1971 EGG HATCHING OF A DESERT FAIRY SHRIMP 47

5- 30,000,umhos 0 day, when 5.6% were in the range from 0.6 to + presence of nauplii I on the -c minimumsize 0.6mm 2.30sI. 4.5 mm. None less than 7 mm were found 2.20%. n minimumsize 3.0mm 2.1%. 70th day, when 97% were 10 mm or longer. On x - * minimumsize 7.0 mm the 108thday 3.6% were 3.0-3.9 mm long. o 1.70 The stimulusfor hatchingafter the 44th day is

W 1.38%. not known, but each of the hatchingscoincided 0 _ / b 1.25%. with meltingof ice. The pond does not freeze 1.0 %. 1 a 2 0.e86% over forany great lengthof time,but duringcold 0 / periodsit is coveredin the morningby a thinsheet E of ice, whichusually melts by noon. In Figure 4

X--W 0 0.48%. io . i are shown the salinitychanges observed on the 1 - - 86th day. The salinitywas at 1400 on Feb- + I 1, 1, ,2I +1 2.1%Yo 00 10 20 30 40 50 60 70 80 go, 108 TIME IN DAYS ruary 19th. At 0700 on the 20th,when the pond FIG. 4. Changes in salinity and presence of young, was completelycovered with ice, the salinityof Winter 1965-66 pond. the unfrozenwater beneathwas 2.3%o. At 1030, just as all the ice had melted,the surfacesalinity ulation) 41% of the shrimpwere mature (10 mm was 1.25%o. Stratificationcontinued until a wind or greater), 38% were 7-10 mm long, 21% 3- came up at noon. By 1400 the salinitythrough- 6.9 mm long. The growth rate of the 10-mm out the pond was 2.5%o. In studiesby others(e.g., individualswas 0.66 mm/day,assuming that they Coopey 1950), snow melt has been reportedas hatched15 days previously. This is Y2 the growth the stimulus for hatchingof branchiopodeggs, rate estimatedfor immaturesof the same species thoughin thesecases the stimuluswas liquid water in summer.B. mackini apparentlytakes two weeks after its absence. In Rabbit Dry Lake liquid to reach maturityin winter. On the 86th day of water was continuallypresent, but the ice-melt the pond the largest individualin the population water could subject eggs to reductionin salinity. was 32 mm long, and a third of the population In this winter pond hatchingwas continuous was longerthan 25 mm. while nearly constant salinities prevailed, but Because naupliiwere stillpresent on the second ceased when salinityincrease. At salinitiesabove of December,continuous hatching can be assumed 1%o hatchingoccurred only coincidentwith dilu- duringat least the firstseven days of the pond. It tion by rainwaterinflow or ice-melt. probablyceased after anotherfive days, because SummerPond, August 2 to 5, 1966.-This was no shrimpless than 3 mm long were presenton a pond of shortduration (66 hrs), of small initial the 15thday (Fig. 4). An initialhatching period volume,and witha continuousand rapid increase of 12 days can thereforebe postulated,during a in salinity. At the end of 42 hours specificcon- period of salinitybelow 1%o and with slight in- ductancewas 5,850 .mhos. crease in salinity. Nauplii collectedon the 27th At 1800 on August 4th, 54 hours fromthe esti- and 44th days showed a long second hatching mated time of origin,as much as possible of the period, which probablystarted with the rain on remainingwater was removed. The three liters the 14th day and continuedfrom the 15thto the collectedcontained a total of 74 individualswith 48th day. Salinity remainedbelow 1%o, and in- a size range of 0.5-1.5 mm. The temperature crease in salinitywas slight. Under these condi- range and the oxygenvalues in the open water of tions hatchingwas continuous,in contrastto pe- this pond were similarto those foundin the sum- riodichatching observed following reductions from mer pond of 1965. And the initialsalinity of ca. highersalinities during periods of increasingsa- 1,000 Vmhoswas the same. Yet in this pond, a linity. It cannotbe determinedwhether the hatch- negligiblenumber of naupliiappeared, considering ing was of dried eggs from previous fillingsof that about a quarter of the basin was initially the pond or of eggs freshlyejected by living fe- coveredand thatthis includedthe lowestpoint in males. From observationsin the laboratory(de- the basin where most of the shrimpcollect at the scribedbelow) of differencesin behaviorof ejected end of each pond, die and decomposeand leave and dried eggs, the eggs most likely to hatch theireggs. The obviousenvironmental differences under conditionsof constantlow salinityare the betweenthis and the pond of the previoussummer ejected eggs-the ones thathave neverbeen dried are the smaller volume and the continuousand nor subjected to low oxygen concentrations. very rapid increasein specificconductance imme- Despite the overall increasein salinityafter the diatelyfollowing formation of the pond. In the 45th day of the pond, nauplii reappeared on the 1965 summerpond in which many eggs hatched, 63rd day, amountingto 6.2% of total population, it took ten days to reach a salinitycorresponding and very small stages were presenton the 86th to 5,800 l.mhos. In this 1966 pond a specificcon- 48 LON R. BROWN AND LARS H. CARPELAN Ecology,Volume 52, No. 1

5 + presence of nouplii 10,800 urmhos Q 5 ,O 20,000 umhos 2.48 %. + presence of nouplii U minimum size 1.0 mm 2.45%. 0 minimumsize 0.6 mm i minimum size 2.0 mm ^ f2.25%. x 4 - / * minimum size 1.0 mm 4 * minimum size 3.0 mm o - 1.72%z1 A minimumsize 4.0 mm 0#A A minimum size 4.0 mm

8 * minimum size 7.0 mm 3 _ <1 .48%. 1.38e 1.56%. 1.2 . 0 1.33 %. 2 2-/ 1.1%. o 2 _ - 1.0%.

-- . o _ oC- 2 0.68e 0 w ~~~~~~~~~~~~~~~~0 a.I 0 W I + I mA I. 0 10 20 30 40 50 60 7T0 e 90 lo 20 30 40 50 60 70 80 90 TIME IN DAYS TIME IN DAYS FIG. 5. Changes in salinityand presence of young,Win- FIG. 6. Changes in salinity and presence of young, ter 1966-67 pond. Spring 1967 pond. ductanceof 5,880 was reachedin 42 hours,a rate 6). The only springpond that occurredduring of increase of more than 2,400 IVmhosper day the studylasted for 28 days. Toward the end of (see Table 3). the pond the increase in salinitywas very rapid. WinterPond, November7, 1966 to February2, On the 23rd,27th, and 28th days the specificcon- 1967. (Fig. 5).-This pond,which lasted 87 days, ductance values were 9,200, 14,000, and 20,000 differedfrom the winterpond of the previousyear Vmhos,corresponding to salinitiesof about 5, 8, in that there was continuousincrease in salinity and 11.5%o. At the last determinationthe depth with only one interruptionby additional water. of the remainingwater was less than 5 mm. On the 29th day the estimateddecrease was from The longestindividual (5.6 mm) on the ninth 2,630 to 2,100lmhos. On the 80th day the spe- day had a growthrate of 0.6 mm per day, assum- cificconductance was 10,800 tmhos,a salinityof ing it hatchedon the firstday. This rateof growth about 6%o. of immaturesis about the same as in winterponds. There were two major periods of hatching-an Less than 1% of the populationon the ninthday initialhatch duringthe first15 days of the pond, was shorterthan 2 mm, which suggeststhat for and a second hatch after the rains on the 29th 99% of the populationhatching had ceased 3-4 day. On the 26th day the smallestindividual was days previously. Accordingto the growthrate, 3 mm long, and 96% of the populationwas over it would be expectedthat 99% of the population 5 mm long. In contrast,on the 32nd day, 92% would be at least 3.8 mm long on the 12th day; of the populationwas less than 3 mm long. The however,on that day 28% of the populationwas secondhatch was essentiallyover by the 35th day, less than3.0 mm long. This changein population with some hatchingup to the 40th day. 6.6% structurebetween the 9th and 12thdays indicates were 3.0-4.9 mm long on the 49th day. A third resumptionof hatchingafter inflow of more water hatchduring the firstweek in January(62nd day) on the 10th day. The presence of nauplii and correlatedwith a period of ice melt. young stages on the 15th to the 19th days coin- There was no dilutionby increase in volume cides witha cold periodwhen ice was on theponds after the 32nd day, as shown by consistentde- in the mornings. crease in depthof the pond. The depthsrecorded SummerPond, September1 to 6, 1967.-This were: 20th day, 5 cm; 29th day, rain; 32nd day, pond lasted 127 hr. Duringthe first28 hr salinity 15 cm; 40th day, 9.5 cm; 49th day, 9 cm; 62nd increasedfrom 1,100 to 2,300 :mhos. On Sep- day, 5 cm; 66th day, 4 cm; 73rd day, 3 cm; 87th tember2 as a resultof a second rain the salinity day, moistmud. Despite decreasingdepth, which decreasedto 1,300 ,umhosin a matterof minutes. is usuallyindicative of evaporationand increasing About 70% of the basin was covered with water. salinity,there were local decreases in salinityon The maximumdepth (in thenortheast corner) was the8th and 12thof January(62nd and 66thdays). 4-5 cm, but most of the water was only 1-2 cm The salinitiesof samples taken just afterthe ice deep. Evaporation was rapid, and during the had meltedwere 1.33 and 1.77%o (Fig. 5). The next 75 hoursthe salinityincreased from 1,300 to occurrenceof small immaturesat timesof ice-melt 5,000 ptmhos. On September5 the water that suggests,as in the pond of the previouswinter, a remainedin the basin was siphonedout, yielding relationbetween lowered salinityand hatching. 24 shrimpranging in size from0.5 to 5.2 mm. Spring Pond, April 11 to May 9, 1967. (Fig. The water was replacedand when resiphonedon Winter 1971 EGG HATCHING OF A DESERT FAIRY SHRIMP 49 TABLE 3. Summary of field hatching periods is not a seasonal species. It seems to be temper- ature-independentin the range occurringon the Salinity Rangeof increase Duration Rateof Mohave desert,appearing whenever there is suffi- salinity during of salinity increase hatching hatching increase cientwater. But periods of egg hatchtook place Date ofpond (jmhos) (pmhos) (days) (jsmhos/day) during lowered salinity,and inhibitionof hatch Winter65-66 ...... 730-960 230 38 6 occurred during increasingsalinity. There was Winter66-67...... 1000-1795 795 15 53 no fieldevidence that oxygen had a role in regu- Winter66-67 ...... 2100-2990 890 12 74 Winter65-66 ...... 1000-1800 800 10 80 latingthe hatchof B. mackinieggs, but the impli- Spring1967 ...... 1000-3450 2450 11-12 213 cation of oxygen as a stimulusand inhibitorof Spring1967 ...... 1500-4250 2750 7 397 hatchof eggs of otherspecies and the presenceof Summer1965 ...... 2600-4000 1400 4-5 315 Summer1965 ...... 1000-3000 2000 2 1000 a gradientof oxygenin the mud surroundingthe Summer1966 ...... 1000-5850 4850 2 2425 eggs in Rabbit Dry Lake made oxygenof interest. Summer1967 ...... 1000-2300 1200 1 1200 Summer1967 ...... 1300-5000 3700 3 1230 Hence, preliminarylaboratory tests were made of the effectsof salinityand oxygenon egg hatch. the sixth (final) day, eight shrimp,the largest Egg collectingand handling 8 mm long,the smallest1.4 mm,were collectedin Dried eggs.-Fifty-gram samples of dry soil the remainingwater. fromthe top 5-10 mm of the study basin were Obviously some environmentalfactor inhibited placed on 80-meshnylon netting. The eggs were the hatching,for only 32 eggs hatcheddespite in- washed free of the fine silt, by a stream of tap undationof 70% of a basin, which at times sup- water, and immediatelydried at room temper- ports 400,000 shrimp. Neither temperaturenor ature. A 50-gram sample of the soil contained lightnor pH nor initialsalinity were differentfrom between 110 and 200 eggs, which at the end of conditionsin the summerpond of 1965, which the washing still had their dehydrated,indented supporteda large population. But relativeto the appearance. It takes the eggs an hour or more 1965 summerpond, the ponds of 1966 and 1967 to hydrate,and the washing took only 3-5 min. had smaller volumes and shorterduration-too Freeing the eggs fromthe mud by washingthem short to allow completionof life cycle. And in even brieflywas not desirable,for it leaves the theseshort-lived ponds few eggs hatched. experimentsopen to the objectionthat as a result of the wetting,the eggs may have been stimulated Duration of hatching to develop. But the washing proved to be the Durationof hatchingis inverselyrelated to rate onlyway to get eggs in knownnumber, uninjured of increasein salinity(Table 3). As the rate of and freeof mud. Althoughthe hatchwas the same increase in specificconductance increased from 6 if the eggs were only blotted,they were allowed to 1,000 'mhos per day, the durationof hatching to redryfor at least 24 hr beforebeing placed in decreasedfrom 38 days to 2. When salinityin- plasticpetri dishes (Falcon Plastics,B-D Labora- creased at rates above 1,000 imhos per day, as in tories, Los Angeles), which were kept in high the summerponds of 1966 and 1967 there was humiditychambers to minimizeevaporation of the virtualinhibition of hatching. testmedia in whichsalinity was the variable. The petridishes were used when oxygenconcentration LABORATORY STUDIES OF EGG HATCH was not controlled. Field observationsreported in the previoussec- Ejected eggs.-Eggs collectedfrom dry soil in tion suggestedthat of the obvious environmental the manner described above are referredto as variablesimmediately surrounding the egg,salinity "dried" eggs in contrastto eggs which are re- is the most likely to affecthatching. Temper- ferredto as "ejected," which were collected di- ature, the most studied of variables,seemed un- rectlyfrom females that had maturedin the field. importantin thiscase. For thoughit was observed Females withfull egg sacs were placed in vials thattemperature influenced growth rate, egg hatch- (Thorton Plastic Co., Salt Lake City) containing ing of B. mackiniwas not inhibitedin the rangeof the testmedium. Afterthe eggs were ejected,the temperatureobserved. Large hatches were ob- femaleswere removedfrom the 40- or 3.5-dram served in winterponds when temperatureswere vial. (Vials of differentsize were used depending 1-140C and in summerponds in the range from on whethera singlebrood or multiplebroods were 12-32TC. There are desert species of Branchi- being tested.) Eggs obtainedin this way were opoda thatare seasonal (e.g., B. gigas) similarto never dried and never subjected to any salt con- whathas been reportedby Dexter (1953), Prophet centrationother than that of the test mediuminto (1963), Moore (1959), and Horne (1967) for which they were ejected. Because the vials are some of the species theystudied. But B. mackini porousto oxygen,tests could be made in themwith 50 LON R. BROWN AND LARS H. CARPELAN Ecology,Volume 52, No. 1 oxygen at equilibriumwith the atmosphere,but TABLE 4. Percentage hatch within 120 hr of dry eggs withoutbubbling with air. In othertests oxygen placed in test solutions concentrationswere controlled by bubblingthrough the mediumeither air, an air-nitrogenmixture, or Specific conductance No. of eggs Total no. Total (,umhos) tested hatched hatch (%) nitrogenalone. Experimentalconditions.-In the non-air-bub- a. Withoutair bubbling.(4 or 7 replicates) bled vials and petridishes, the oxygen concentra- tion in the "macroenvironment"determined by 360 699 619 88.4 polarographicprobe was always above 7 mg/i. 1,000 649 528 81.4 "Macroenvironment"is emphasized,because there 5,000 897 4061 45.3 10,000 827 71 0.86 is, by inferencefrom other studies (Table 5), a 20,000 842 21 0.24 lower oxygen concentrationat the bottomof the dish or vial in the immediatevicinity of the egg b With air bubbling(2 or 5 replicates) surface. The concentrationsat the egg surface ("microenvironment")become of concern when 360 352 233 66.2 eggs are allowed to rest at the bottomof a vessel 1,000 368 160 43.5 withoutagitation, simulating the natural condi- 2,000 280 109 39.0 3,000 234 80 34.2 tions in which eggs lie at the bottomof a pond. 4,000 279 77 27.6 In other tests, under less natural conditions,the 5,000 705 232 3.3 eggs were kept in constantmotion by bubbling 17 replicates withgas. In thesetests the oxygenconcentration 25 replicates at the egg surfacecould be assumed to be at the concentrationmeasured in the "macroenviron- in salinityas well as upon the initialsalinity from ment." whichthe dilutionwas made (Fig. 7). There is no absolutethreshold below whichhatching occurs Results but with inhibitionabove that salinity. Rather it General hatchingcharacteristics.-Hatching of is a reduction,a change relativeto the initialcon- dried eggs beginswithin 24 to 36 hr afterwetting centration,that serves as stimulus. At 3,000 or aftera substantialreduction in salinity. After Emhosa 10% dilutionresults in about 50% hatch. terminationof a hatch,the remainingunhatched At salinitiesabove 3,000 Vmhosa greaterdilution eggs do not hatch unless there is anotherreduc- is requiredto obtain 50% hatch. Thus at 6,000 tion in salinity. In contrast,the response of and 9,000 Vmhosat 50% dilutionresults in slightly ejected eggs is less predictable,liable to either morethan a 50% hatch; at 70,000 Vmhosan 80% bursts of hatchingor continuoushatching at all dilutionproduces 10% hatch. The resultingdilu- salinitiesand apparentlyindependent of changes tions,the concentrations at whichthe eggs hatched, in salinity,perhaps more dependentupon oxygen were 2,000, 3,000, 4,500 and 14,000 pmhos. So concentrations(see below). the eggs are inhibitedfrom hatching at constant Effectof salinityon hatching.-In the follow- salinities,which may be lower than concentrations ing experimentsthe eggs were on the bottomof at whicheggs will hatch,providing that the higher plastic vessels in which the oxygen concentration salinityis a reductionfrom a still higherconcen- of the mediumwas at saturationwith air but with- tration. That is, hatchingresults from a change, out the agitationprovided by bubbling. Salinity a stimulus. was estimatedfrom specificconductance of 2-ml b. Ejected eggs.-In contrastto driedeggs, some samples drawn into a pipettecell. Salinityat the ejected eggs afterejection into water from the egg egg surfacewas assumedto be the same as thatof sac will hatch thoughthey have never been sub- the medium. jected to reductionin salinityand have neverbeen a. Dried eggs.-With increase in salinitypro- dried. The proportionthat hatches depends on gressivelyfewer eggs hatchwhen they go fromdry the salinity,though only on the average because to the wet state (Table 4a). Whereas morethan of great variabilityamong the eggs, even those of 80% of eggs hatch at 360 and 1,000 Cmhos,only a single brood. Eggs ejected into water at 1,000 halfof themhatch at 5,000 Vmhos. At stillhigher tumhosbegan to hatchtwo days afterejection, with concentrations(10,000 and 20,000 jmhos) less 17-73% of the eggs in each broodhatching during than 1% hatch. In only one of seven trials at a 5-day period. In a mixtureof broods (eggs 20,000 was thereany hatch at all; in that sample from several females), 585 of a total of 1,742 two of 149 eggs hatched. (33%) hatchedduring a 9-dayperiod afterhatch- Upon dilution,by additionof water to the me- ing began four days after ejection. At this sa- dium, responsedepends on the relativereduction linity80% of dried eggs hatch. At 2,600 ptmhos Winter 1971 EGG HATCHING OF A DESERT FAIRY SHRIMP 51 hatchwhen the oxygen of the mediumis in equi- libriumwith the air. Four replicatesof driedeggs (387, 178, 144, and 164) totaling873 eggs were 80 - testedin water of this salinity(360 pmhos), but TO-~~~~~~~ with oxygen maintainedat less than 1 mg/1 by an air-nitrogenmixture. There was completein- hibition:at the end of 120 hr no eggs had hatched. To verifyviability, three of the four vials were then bubbled with air. In them 60-70% of the eggs hatchedduring 120 hr. In the fourthvial, whichwas continuedas a controlwith oxygen at 7060 0.1-1.0 mg/l,inhibition of hatchingcontinued. Similar tests were made with eggs that had been stored wet at 20,000 ,mhos with oxygen in equilibriumwith air. On transferto 360 pmhos water (also withoxygen at saturation)there was 80-90% hatchduring a 120-hrperiod. On trans- ferfrom 20,000 pmhos (and withoxygen in equi- libriumwith air) to 360 'pmhosbut with oxygen 0, 2404 )0 0 0 at 0.1-1.0 mg/1,there was completeinhibition of hatching. Subsequentair-bubbling produced a 60- 0 ',0 umhos 1| 70% hatchin 120 hr. In thesetests in watersof low salinity,eggs were inhibitedfrom hatching at oxygen concentrations below 1 mg/1. It seemsthat at low salinitynearly completeoxygen depletionis needed for inhibi- tion: slightlyhigher concentrations of 1-2 mg/l umhoss at~days.~ Inasnl7~~ 3,00~umho. rodsuya 0,000 did not inhibithatching. The observationthat only 60-70% hatch in agitatedsolutions, as com- pared to an 80-90% hatch in nonagitatedsolu- hatchwas a funciong12diluioof period. FI.vererenage tions, showed the possibilitythat noninhibiting Inicotialsalntyidriaed.eg r niie rmhthn dayinof tssie 360Eumhos b intermediateconcentrations of oxygen at the bot- tom of nonagitatedvials mightenhance hatching thanedredeggs bfrbeganin to aftehatch threeas nd and be responsiblefor the differencesin hatchob- 365ontnu1to32 hatchfrlned durings Theuetsevena served in agitatedand nonagitatedtest solutions. water0,0009o0thio sa Tests were thereforemade with agitationpro- evercfondhatchiongFIG.7.terentage bgin four hat action pond),of ejctondilu and ftheothe o Intialasalntbgin iondicted. fosurdaysejestdays aiite after, ejcton ih nuhttonger vided by bubblingwith air. Comparisonof re- avragedegofejected hatchedeggs beganrewaithduingtw toaticiaf daysto hatchdAbdy period.l and sults shown in Table 4b with those in Table 4a, 3o65Inmcontrat ofte1,573drprtied(775)o () hrarejebteocnedr eggsprions roThenet actdintoseean shows that at all salinities tested fewer eggs ays.o ta satigle rpoostility that100e0 than hatchedwhen kept in motionin air-bubbledsolu- tionsthan when the eggs lay quietlyat the bottom. At 5,000 Vmhosvery few eggs hatchedwhile air- tha driced beforebeginninge to hatchbutte bubbled. This was surprising,because 40-50% averageof eggs, wass 1%n hatchding1- of contneotoe hactcforlneongerpeiods. Thunstintsa- the eggs had hatchedwithin 120 hr in the tests tabulated in Table 4a. In five trials with air- pairitoftessat 360ludpmoss(abslinty lowcanerthn bubblingat 5,000,mhos theaverage hatch was only 3.3%: (6/118; 2/116; 2/141; 0/163; 13/167). smallert kis- o in cnthrast,prosmorticson driedeggs dared tegg inhibit fejectedlus. hatching These data suggest that a transitoryperiod of drie eggsd twcedi ays. Af slgtl haedgswithin low oxygen at the egg surface may have been presentin the nonagitatedtests and that this pre- vious exposure might account for the enhanced hatchin nonagitatedtest media, especially at 5,000 ,umhos. Thereforeanother test was made at 5,000 ,umhoswith oxygenmaintained below 2 mg/1 for 120 hr. There was no hatching. The testmedium salinitypmhos),(360 in whichthe eggs are free was then bubbled with air, and during the next frominhibiting effects of salinity,there is 80-90d 120 hr 47-68% of the eggs hatchedin threerepli- 52 LON R. BROWN AND LARS H. CARPELAN Ecology,Volume 52, No. 1 cates totaling544 eggs. This is far greaterthan tory culturesnon-dried eggs seem to hatch with- the 3% hatchobserved with air-bubblingwithout out stimulationby reductionin salinity. It seems a priorperiod of low oxygenand is in the range likelythat continuoushatching probably does not foundfor nonagitated solutions (Table 4a). occur in Rabbit Dry Lake, because salinitythere These resultsshow that a period of low oxygen is seldom constant,especially during later stages tension does enhance hatch. However, since a when salinityis high. Arrest of developmentof good proportionof eggs will hatchwithout it, ex- both non-driedand dried eggs is probably the posure to low oxygen is not mandatory,as has resultof increasingsalinity (Table 3). But there been found for some species of arthropodeggs is the possibilitythat the non-driedeggs are also (see Discussion). Enhancementof hatchby pre- inhibitedby low oxygenat the bottom,as reported vious exposure to low oxygen tensionis less no- by Moore (1963) in waters withoutmarked sa- ticeableat low than at high salinities(Table 4b), linitychanges. and the effectof oxygen as an inhibitorof hatch The effectof bottomsediments on theirhatching also varies withsalinity. At 360 and 1,000,mhos, was investigatedin tests on eggs ejected into sa- hatchingis inhibitedonly by concentrationsbelow linities from 1,000 to 10,000 ;Vmhos. At each 1 mg/1,and thereis no inhibitionat high concen- salinityabout 10,000 eggs were collected (from trations. At 5,000 pmhosthere is inhibitionboth 100 mature females each with about 100 eggs). below 2 mg/1, and above 7 mg/1,provided that When egg collectionwas complete,the eggs were no exposureto low oxygenconcentrations has oc- transferredto water of the same salinitythat had curredwhile eggs were hydrated. been centrifugedto get rid of the suspendedsoil. b. Ejected eggs.-When eggs were ejected from Following removalof sediments,the oxygencon- the ovisac of livingfemales into a constantlyair- centrationwas above 7 mg/1 in all the solutions. bubbled solution of low salinity (360 Vmhos), Within 1-5 days hatchingbegan at all salinities. hatchingbegan in 36-48 hr insteadof the 120 hr About 1,000 eggs were removedas controlsand thatwas characteristicof nonbubbledmedia, indi- kept in sediment-free(centrifuged) water, and catingdelay of hatchingby the lower oxygensur- then, to simulate field conditions,just enough roundingthe ejected egg at the bottom of the sedimentwas added to the vial to cover the re- nonagitatedmedium. There was delay, rather mainderof the eggs. Resultsin Table 5 show that than inhibition,because eventuallythe same pro- inhibitionof hatchingcould occur, except during portionof eggs hatchedin the nonagitatedsolution the first24 hr, even thoughthe eggs were covered (67-75%) as in the agitatedsolution (6046%). by no more than 1-2 mm of sediment. Hatching In contrastdried eggs at 360 ,mhos did not show continuedin the vials withoutsediment. The in- a time differencein initiationof hatching,but the hibitionof hatchingoccurred at intermediateox- finalhatch was 60-70% in the air-agitatedmedium ygen tensions (2.7 mg/1 and less) as measured versus 80-90% in the nonagitatedmedium. withthe oxygenprobe in the bottom5 mm of the The delayof hatchingof ejectedeggs by lowered vials. Althoughthe nauplii presentat 1,000 and oxygenconcentration offered an additionalexpla- 2,000 Vmhosmay be too few to be significant,it nationfor apparent inhibition of hatchingof ejected could be thatat low salinitiesonly very low oxygen eggs in the long-lastingwinter ponds. In the concentrations(near depletion) will inhibitthe laboratorycultures, once a femaleejects eggs, she hatchof ejected eggs,as was foundfor dried eggs. producesanother clutch, is inseminatedand again In the centrifugedwater, where oxygen was ejects eggs in 2- to 4-day cyclesuntil she dies or not reducedbelow 5.6 mg/1,the proportionhatch- the pond dries up. At constantsalinity in labora- ing (2-10%) seems low, but this does not include

TABLE, 5. Inhibition of hatching by natural sediments

Salinity(~tmhos) Oxygentension after 6 days (mg/i) No. hatchedin 6 days No./ca. 1000 No./ca. 8000 Sediment-free With sediments eggs eggs Sediment-free With sediments top/bottom top/bottom sediment-free with sediments 500 500 7.0/6.5 6.1/2.7 100 0 1000 1000 6.3/6.0 6.1/2.0 72 4 2000 2000 7.0/6.6 6.6/2.0 65 12 3000 3000 7.1/6.9 6.5/1.9 63 0 4000 4000 7.4/7.0 6.5/1.5 59 0 5000 5000 7.5/7.5 6.4/2.7 38 0 6000 6000 7.4/7.3 5.4/2.3 74 0 7000 7000 7.0/5.6 6.2/1.4 19 0 Winter 1971 EGG HATCHING OF A DESERT FAIRY SHRIMP 53 those that hatchedin the firstpart of the experi- 6,000 and 9,000 :pmhosa 50% hatch requires a ment. 35-40% dilution. Eggs will hatch at high salin- The oxygenconcentrations measured at the bot- ities, but a great dilutionis requiredfor a small tom of the vials are in the range surroundingthe proportionof the eggs to hatch. There seems to eggs at the bottomof the pond. It is thuspossible be no thresholdvalue below whichhatching always that low oxygenmay preventcontinuous hatching occurs,with absolute inhibition above thatsalinity; of ejected eggs except at low salinities,and that ratherit is a change,a reductionin concentration, oxygenplays a role in inhibitinghatch as well as that serves as stimulus. in conditioningeggs to respondto salinitychanges. The response of ejected eggs-those never dried-is less predictable,and some hatchwithout DISCUSSION reductionin salinity. Ratherthan needing a stim- In reviewinglife histories of freshwater Branchi- ulus to hatch,non-dried eggs seem to require an inhibitorto prevent opoda, Hutchinson(1967, p. 557) statesthat they hatching. Low concentrations "live in watersthat are subject to periodicdrying, of oxygenin the bottommud seemto providesuch or extensive freezing,or, as in the habitats of inhibition. Artemia,are excessivelysaline." This reportcon- Though the "fundamental"stimulus for hatch- ing may, cerns a species living in temporaryponds in an indeed,be loweredosmotic pressure, the arid regionwhere all threefactors-drying, freez- regulationof developmentis complex. In a humid- ing, and salinitychanges-are involved,though region study Broch (1965) found that hatching the freezingin not extensiveand the salinityis of Chirocephalopsisbundyi was initiatedby low not excessive. Hutchinsonfurther suggests (p. oxygenconcentrations, but thatdevelopment could be 566) that "the main differencebetween species of arrestedby other factorsat any one of four humidand arid regions,if there is such a differ- stages if conditions were unfavorablefor con- ence, is most likelyto lie in the facultativenature tinued development. For example, one of the phases of the diapause in those species . . . that inhabit could be stopped fromdeveloping by low humidregions." This implicationthat an obligate temperatureand/or anaerobic conditions. The diapause exists in desertspecies is most unlikely latteris similar to controlof developmentof B. because of the year-to-yearvariability in time and mcackiniby low oxygenin the bottommud. Ability amount of rain. It seems more likely that the to stop developmentis an importantadaptation, dormantstage always respondsto an outsidestim- because if development,once started,were irre- ulus and thatthe diapause thereforeis facultative. vocable,synchronization of developmentwith con- In his review of the various factorsand com- ditionsfavorable for completion of lifecycle would binationsof factors that have been reportedto not be possible. stimulateand inhibithatching, Hutchinson states Experimentson the effectof oxygenconcentra- tions on (p. 562) "it is quite likelythat the fundamental the dried eggs of B. mackinishowed that change needed to produce hatchingis always a the effectsvaried with salinity. The two factors loweringof osmoticpressure at the egg surface." interact. Inhibitionof hatchingoccurs at all sa- This opinionof Hutchinsonis essentiallycorrob- linitieswhen oxygentension is less than 1 mg/1 orated by observationsin the desert pond, and (at 20-250C). The oxygenconcentration required verifiedin the laboratory,but the regulationof forhatching of previouslydry eggs increasesfrom developmentis not that simple,and it was found 1 mg/1at 360 pmhos,to 2 mg/1at 5,000 pmhos; that dried and non-driedeggs behaveddifferently. at 10,000pmhos and above feweggs hatchregard- Hatching of dried eggs of B. wackini begins less of oxygentension. With increasein salinity, within24-36 hr afterwetting or followinga sub- high oxygen concentrations(7-9 mg/1) become stantialreduction in salinity. In going fromthe increasinglyinhibitory to dried eggs, unless the dry to the wet state, progressivelyfewer eggs eggs, after hydration,were previously "condi- hatch at higher salinities. Thus, whereas 80% tioned" by low oxygenconcentration. With con- of dried eggs hatched when placed in waters at stantlyhigh oxygen concentrations there is a sharp concentrationsbelow 1,000pmhos, only 50%o hatch rise in inhibitionof hatchingat salinitiesabove at 5,000 ,umhos,and fewerthan 1% hatch when 4,000 pmhos. placed in waters more concentratedthan 10,000 Dried and non-driedeggs differin responseto pWmhos.After the initialhatch, response to dilu- oxygenwhen in highersalinities (10,000 pmhos). tion depends both upon the relativereduction in Non-driedeggs are stimulatedto hatchby changes salinityand upon the salinityprior to dilution. in oxygen concentration,a response shown by At the relativelylow salinityof 3,000 pmhos, a dried eggs only at low salinities. With increase 10% dilution results in a 50%o hatch, while at in salinity,oxygen has a decreasingrole in the 54 LON R. BROWN AND LARS H. CARPELAN Ecology,Volume 52, No. 1 regulationof hatchingof driedeggs. The delayed LITERATURE CITED hatch and prolongedperiod of hatch of ejected Bishop, J. A. 1967. Some adaptations of Limnadia (non-dried)eggs, is eliminatedby preconditioning stanleyana King (Crustacea: : Concho- straca) to a temporary fresh water environment. J. at low oxygen concentration. After such treat- Ecol. 36: 599-609. ment.dried and non-driedeggs respondalike. Bouldin, D. R. 1968. Models for describing the diffu- It seems probabletherefore that the differences sion of oxygen and other mobile constituents across betweenspecies of humidand arid regionslie not the mud-water interface. J. Ecol. 56: 77-87. in differenttypes of diapause but in differentre- Brewer, R. H. 1964. The phenology of Diaptomus stagnalis (Copepoda: Calanoida): The developmentand sponsesto changesin salinityand oxygenconcen- hatching of the egg stage. Physiol. Zool. 37: 1-20. trationsand to the degree of such changes. The Broch, E. S. 1965. Mechanism of adaptation of the osmoticchanges in freshwatersof humid regions fairy shrimp Chirocephalopsis bundyi Forbes to a tem- are so slightthat studentsof freshwaterbranchi- porary pond. Cornell Exp. Sta. Memoir 392: 3-48. opods generallydisregard salinity, justifiably con- . 1969. The osmotic adaptation of the fairy shrimp Branchinecta campestris Lynch to saline astatic waters. sideringit to be a factorof no concern. Thus Limnol. Oceanogr. 14: 485-492. neitherBroch (1965) in his studyof Chirocepha- Clegg, J. S. 1964. The control of emergence and me- lopsis bundyi,nor Moore (1963) workingwith tabolism by external osmotic pressure and the role Streptocephalusseali and Eubranchipusholmani, of free glycerol in developing cysts of Artemia salina. nor Bishop (1967) withLimnadia stanleyanaeven J. Exp. Biol. 41: 879-892. Cole, G. A., and R. J. Brown. 1967. The chemistry of mentionsalinity. Their studies showed low ox- Artemia habitats. Ecology 48: 858-861. ygen concentrationsto be the factorstimulating Coopey, R. W. 1950. The life history of the fairy hatchingof some species and inhibitinghatching shrimp Eubranchipus oregonus. Amer. Microsc. Soc. of others. It seems logical that in humidregions Trans. 69: 125-132. some factor other than water would Dexter, R. W. 1956. A new fairy shrimp from western stimulate U. S., with notes on other North American species. hatchingof eggs that are constantlymoist. Or J. Washington Acad. Sci. 46: 159-165. lookingat it fromanother point of view, some fac- . 1959. Anostraca. p. 558-571. In Edmondston, tor mustchange to removean inhibitionof further W. T. (ed.) Freshwater Biology, 2nd ed. Wiley, New development. It is not unexpectedthat the rel- York. 1248 pp. Dexter, R. W., and M. S. Ferguson. 1943. Life history ativelygreat changes in salinitythat occur in arid and distributional studies on Eubranchipus serratus regions can serve to correlatehatching with the Forbes. Amer. Midl. Nat. 29: 210-222. likelihood of completinga life cycle. Osmotic Gjullin, C. M., C. P. Hegarty, and W. B. Bollen. 1941. pressure could act to inhibithatching (at high The necessity of a low oxygen concentration for and stimulate hatching of Aedes mosquito eggs. J. Cell. Comp. salinity) to hatching (at low sa- Physiol. 17: 193-202. linity). Horne, F. R. 1967. Effects of physical-chemical factors In both humidand arid regionsthe stimulation on the distribution and occurrence of some south- of the egg to hatch must be by some factorthat eastern Wyoming phyllopods. Ecology 48: 472-477. changes at the time of originof the pond. Tem- Horsfall, W. R. 1956. Eggs of floodwater mosquitoes. III. Conditioning and hatching of Aedes mosquito peraturecan servethis functionin ponds originat- eggs. Ent. Soc. Amer. Ann. 49: 66-72. ing from snow melt. As Hutchinson (p. 5655) . 1958. Eggs of floodwater mosquitoes. V. Ef- says of Conchostraca,"those species that inhabit fect of oxygen on hatching. Ent. Soc. Amer. Ann. humidregions have life 51: 209-213. generally cyclesdetermined Hutchinson, G. E. 1967. A Treatise on Limnology, mainlyby temperature." But if the water in the Vol. 2. John Wiley and Sons, Inc. ix, 1115 p. basin undergoessignificant changes in salinity,as Jennings, R. H., and D. M. Whitaker. 1941. The effect it does in arid regions,control of egg hatch may of salinity upon the rate of excystment of Artemia salina. Biol. Bull. (Woods Hole) 80: 194-201. be independentof temperatureand dependentupon Moore, W. G. 1963. Some interspecies relationships bothsalinity and oxygen. The differenteffects of in Anostraca populations of certain Louisiana ponds. various combinationsof these two factorsprovide Ecology 44: 131-139. means for stimulatingand inhibitingdevelopment. Prophet, C. W. 1963. Physical-chemical characteris- tics of habitats and seasonal occurrence of some The eggs themselvesare highlyvariable in re- anostraca in Oklahoma and Kansas. Ecology 44: sponse,and the presenceof at least two types of 798-801. eggs, capable of respondingdifferently as the re- Rzoska, J. 1961. Observations on tropical rainpools sult of differenthistories and general remarks on temporary waters. Hydro- (conditioning)increases biologia 14: 265-286. variability. The heterogeneousenvironment pro- Sussman, A. S., and H. 0. Halvorson. 1966. Spores: vided by an oxygengradient in the mud surround- Their Dormancy and Germination. Harper and Row, ing the eggs providesadded insurancethat all eggs N. Y. 354 p. will at Weissman-Strum, A., and S. H. Kindler. 1963. Hatch- not hatch once underany single set of con- ing of Aedes aegypti (L.) eggs, a two-stage mech- ditions. anism. J. Insect Physiol. 9: 839-947.