PHYSIOLOGICAL ECOLOGY
Embryonic Developmental Rates of Northern Grasshoppers
(Orthoptera: Acrididae): Implications for Climate Change
and Habitat Management
DENNIS J. FIELDING1
AND
LINDA S. DEFOLIART
USDAÐARS, Subarctic Agricultural Research Unit, PO Box 757200, Fairbanks, AK 99775
Environ. Entomol. 39(5): 1643Ð1651 (2010); DOI: 10.1603/EN09356
ABSTRACT Accurate models of temperature-dependent embryonic developmental rates are important to assess the effects of a changing climate on insect life cycles and to suggest methods of
population management by habitat manipulation. Embryonic development determines the life cycle
of many species of grasshoppers, which, in cold climates, spend two winters in the egg stage. Increasing
temperatures associated with climate change in the subarctic could potentiate a switch to a univoltine
life cycle. However, egg hatch could be delayed by maintaining a closed vegetative canopy, which
would lower soil temperatures by shading the soil surface. Prediapause and postdiapause embryonic
developmental rates were measured in the laboratory over a wide range of temperatures for Melanoplus borealis Fieber and Melanoplus sanguinipes F. (Orthoptera: Acrididae) A model was Þt to the
data and used to predict dates of egg hatch in the spring and prediapause development in the fall under
different temperature regimens. Actual soil temperatures were recorded at several locations over 5
yr. To simulate climate warming, 2, 3, or 4⬚C was added to each hourly recorded temperature. Results
suggest that a 2, 3, or 4⬚C increase in soil temperatures will result in eggs hatching ⬇3, 5, or 7 d earlier,
respectively. An increase of 3⬚C would be required to advance prediapause development enough to
allow for a portion of the population to be univoltine in warmer years. To simulate shading, 2 and 4⬚C
were subtracted from observed temperatures. A 4⬚C decrease in temperatures could potentially delay
hatch by 8 d.
KEY WORDS Orthoptera, Acrididae, insect life cycles, voltinism, global warming
Many insects living in cold climates have a multi-year
life cycle (Alexander 1973, Mikkola 1976, Heliovaara
and Vaisanen 1984, Kukal and Kevan 1987), including
some species of grasshoppers which spend two winters
in the egg stage (Kreasky 1960, Fielding 2008). The life
cycle of these grasshoppers is determined in large part
by temperature-dependent rates of embryonic development (Fielding 2006, 2008). Grasshoppers in regions with a short growing season oviposit late in the
summer when cool soil temperatures permit only limited morphological development during the remainder of the season. During the following summer,
embryonic development proceeds until diapause
intervenes. Diapause is terminated during the ensuing
winter, and hatching occurs in the second summer. In
warmer climates where grasshoppers are mostly univoltine, soil temperatures are warm enough for embryos to reach diapause during the Þrst summer and
hatching can occur the following summer. Furthermore, in some populations where warm seasons are
long enough for a complete life cycle in 1 yr, eggs may
hatch after overwintering regardless of the stage of
development attained by the embryos (Church and
1
Corresponding author, e-mail: dennis.Þelding@ars.usda.gov.
Salt 1952, Cherrill and Begon 1991, Fielding 2008).
This avoidance of diapause ensures a univoltine life
cycle even in eggs produced too late in the season to
attain diapause stage before winter (Fielding 2006).
As climate change proceeds and growing seasons
lengthen, the cline from multivoltine to univoltine to
semivoltine (biennial) life cycles of insects may be
expected to shift to higher latitudes and altitudes (Morimoto et al. 1998, Musolin 2007, Tobin et al. 2008,
Altermatt 2010, Martin-Vertedor et al. 2010). In the
case of grasshoppers where oviposition begins shortly
after adult eclosion and continues as long as adults
survive or until cool temperatures in the fall curtail
oviposition, the population of overwintering embryos
will be in various stages of development, depending on
when the eggs were produced (Gage et al. 1976).
Thus, warmer soil temperatures associated with climate change in the subarctic may permit more advanced development before the onset of winter and
could allow for a portion of the population to attain a
univoltine life cycle. Surface air temperatures in the
summer in Alaska are projected to increase by 2Ð 4⬚C
before the end of this century (Christensen et al.
2007). An assessment of the effects of climate change
on grasshopper life cycles therefore depends in large
0046-225X/10/1643Ð1651$04.00/0 䉷 2010 Entomological Society of America
1644
ENVIRONMENTAL ENTOMOLOGY
part on accurate temperatureÐ developmental rate
functions.
Rates of embryonic development are also germane
to grasshopper pest management through habitat manipulation. Vegetation height and growth form can
alter the microclimate experienced by grasshoppers
with consequences for subsequent population dynamics (Branson et al. 2006). For instance, grazing by
livestock can increase temperatures in the soil where
grasshopper eggs are deposited (OÕNeill et al. 2003).
Increased shading by greater canopy coverage by
shrubs or a thick layer of plant litter may reduce soil
temperatures, delaying hatch (Pierson and Wight
1991). Grasshopper hatching phenology relative to
host plant availability and quality is an important component of Þtness (Mukerji et al. 1977). Experiments by
Pickford (1960) in Saskatchewan showed that M. sanguinipes that hatched later in the season had reduced
nymphal survival and adult fecundity.
The objective of previous studies of embryonic development and grasshopper phenology typically was
to predict the timing of egg hatch (Gage et al. 1976,
Kemp and Sanchez 1987) as an aid for scheduling
sampling and control activities. The objective of this
study is to predict the effect of altered temperature
regimens, such as may result from global climate
change or different vegetative covers, on the life cycles of grasshoppers in high-latitude populations. Two
common crop pests from the subarctic, Melanoplus
borealis Fieber and M. sanguinipes F. (Orthoptera:
Acrididae), were studied, and for comparative purposes, a population of M. sanguinipes from Idaho, a
region with a longer growing season. Egg pods obtained from laboratory-reared grasshoppers were incubated at various temperatures to determine rates of
prediapause (from ovipostion to diapause) and postdiapause development (from diapause completion to
hatch). Temperature-developmental rate functions
were Þtted to the data and used to estimate embryological development in the Þeld using soil temperatures recorded over several years.
Materials and Methods
Population Origins, Rearing Methods. Eggs were
obtained from laboratory colonies of M. borealis and M.
sanguinipes. The Idaho population of M. sanguinipes was
initiated with at least 200 individuals collected near
Lewiston, ID (46.38⬚ N, 117.02⬚ W, 450-m elevation) as
fourth and Þfth instars in 2003 and 2004. Laboratory
colonies of M. sanguinipes and M. borealis from Alaska
were initiated with similar numbers of insects collected
near Delta Junction, AK (64.00⬚ N, 145.73⬚ W, 400-m
elevation). Grasshoppers were maintained on romaine
lettuce and wheat bran at room temperature with incandescent lights (16-h photophase) suspended above
the cages to allow the grasshoppers to self-regulate their
internal temperatures. Plastic cups (500 ml) of moist
sand were provided for oviposition. Offspring from the
Þeld-collected grasshoppers were reared under the same
conditions described above, and all experiments were
conducted on F2 and F3 generation eggs to avoid ma-
Vol. 39, no. 5
ternal effects from the Þeld-collected generation. Oviposition cups were sifted daily to obtain egg pods of
known age. Egg pods were stored in moistened vermiculite in small plastic cups (50 ml) with perforated lids.
Populations were reared separately to avoid outcrossing,
and we assumed no artiÞcial selection in the laboratory
environment had occurred within the one or two
generations.
Eggs were incubated in controlled environment
chambers with temperature control of ⫾0.5⬚C. Preand postdiapause developmental rates (1/d) were determined at Þve constant temperatures from 18 to 30⬚C
at 3⬚ intervals. Some development may occur at temperatures too low to permit hatching, or in the case of
prediapause development, rates at low temperatures
may be nonzero, but be so slow as to make multiple
experiments impractical. To estimate developmental
rates at such low temperatures, four diurnally alternating temperature regimens (27:9, 27:12, 27:15, and
27:18⬚C, each with 12:12-h thermoperiods) were used.
Developmental rates at these lower temperatures
were estimated as the rate observed under the alternating temperatures minus one half the known daily
rate of development at 27⬚C. For example, if there
were no development occurring during the low temperature phase of the diurnal thermoperiod, the time
it takes for an embryo to go from oviposition to diapause at an alternating temperature regimen would be
double the time required at 27⬚C.
Postdiapause Development. Eggs pods were maintained at 24 ⫾ 1⬚C for 40 d to ensure that the embryos
were in diapause. These egg pods were transferred to
5⬚C during diapause development. After at least 90 d
at 5⬚C, egg pods were transferred to one of the nine
experimental temperatures and inspected daily until
hatching ceased (or after 45 d if no hatching occurred). The number of days to hatch was determined
in at least 125 eggs at each of the temperature regimens
described above. Logistic regression was used to generate estimates of median time to hatch. After 5 d
without any new hatches, any remaining unhatched
eggs were examined to determine whether the egg was
viable (fully formed embryo), and percentage of viable, unhatched eggs was calculated.
Prediapause Development. The timing of diapause
induction was determined by measuring CO2 production by the embryos at each of the temperature treatments. As an insect embryo enters diapause, respiration rate drops dramatically, providing a reliable
indicator of diapause status (Gray et al. 1991). Eggs
were carefully removed from pods with the chorion
intact and placed in 5-ml syringes equipped with stopcocks. Three milliliters of room air, which had passed
through a soda-lime column to remove CO2 and then
through distilled water to prevent drying of the eggs,
was drawn into the syringes, which were sealed. Syringes were incubated at 22 ⫾ 1⬚C for 18 Ð24 h. For
every nine syringes with eggs, one empty syringe was
included as a control. CO2 was measured with an
infrared gas analyzer (LI-6252; LICOR, Lincoln, NE).
Two milliliters of air from the syringes was injected
into a CO2-free airstream and passed through a
October 2010
FIELDING AND DEFOLIART: EMBRYONIC DEVELOPMENTAL RATES IN GRASSHOPPERS
magnesium perchlorate column to remove water before passing through the gas analyzer. Flow rate
was maintained at 100 ml/min STP with a mass ßow
controller (C100 SmartTrak; Sierra Instruments,
Monterey, CA). Analog signals from the CO2 analyzer
were converted to digital output (ppm) and recorded
with a computer running data acquisition software
(Sable Systems, Las Vegas, NV). By integrating the
concentration of CO2 in the airstream against time, the
total milliters of CO2 in the 2-ml air sample was
determined. CO2 produced by the egg was calculated
by subtracting the volume of CO2 measured from the
empty syringe (⬍10% of that found in syringes with
eggs) and correcting for the total volume of the syringe while incubating (3 ml). Results were converted
to microliters CO2 per hour per egg.
Respiration rates were measured in samples of
18 Ð30 eggs at 2-d intervals as they approached diapause. New eggs were used at each sampling, i.e.,
repeat measurements were not made on individual
eggs. Previous studies showed that CO2 production
declined from a peak of ⬇0.4 Ð 0.5 ␮l/h/egg (for M.
sanguinipes and M. borealis, respectively) shortly before diapause, to ⬇0.1 ␮l/h/egg during diapause
(Fielding 2008). For both populations of M. sanguinipes, a rate of 0.30 ␮l CO2/h/egg was used to indicate
the beginning of diapause. A higher rate, 0.35 ␮l CO2/
h/egg, was used for M. borealis because previous studies indicated that eggs of this species had higher respiration rates (Fielding 2008). The time required for
embryos to reach the diapause stage was calculated
from the linear regression of respiration rate on age,
from the time of peak respiration rate until respiration
declined to diapause levels (Proc GLM, SAS Institute,
2003). ConÞdence limits for the estimates of the age
at which diapause begins were calculated using the
method described in Zar (1999) for calculating inverse
conÞdence limits.
After respiration rates had dropped to diapause levels, several embryos from each temperature treatment
were examined to verify the stage of development at
diapause. QuantiÞable morphological signposts are
difÞcult to discern at later stages of development, but
the length of the hind femur in relation to the abdominal segments is an indicator of embryonic development (Slifer 1932, Moore 1948, Salt 1949, Salzen
1960). For each embryo examined, the nearest onehalf abdominal segment to which the distal end of the
hind femur reached was recorded.
Model Fitting. A function developed by Logan et al.
(1976), and modiÞed by Lactin et al. (1995), was used
to model developmental rates (1/d, where d is the
number of days required to reach a given developmental stage) as a function of temperature. The modiÞcation of LoganÕs original function by Lactin et al.
(1995) allows for two equations relating development
to temperature. The Þrst is
r共T兲 ⫽ e 共␳*T兲 ⫺ e 共␳*Tmax ⫺ 共Tmax ⫺ T兲/⌬兲
[1]
where T is temperature, r(T) is the temperature-speciÞc rate of development, and ␳, Tmax, and ⌬ are parameters to be Þt to the data. This Þrst equation has no
1645
threshold of development at low temperatures but
only approaches zero asymptotically. The second
equation is similar to the Þrst
r共T兲 ⫽ e 共␳*T兲 ⫺ e 共␳*Tmax ⫺ 共Tmax ⫺ T兲/⌬兲 ⫺ ␭
[2]
but with the addition of a constant, ␭, that allows for
rates of development of zero or less. In practice, predicted negative developmental rates were set to zero.
Both equations were Þt to the pre- and postdiapause
data using maximum likelihood methods (Proc model,
SAS Institute 2003). The better Þtting model for each
data set, as determined by log likelihood and adjusted
r2, was used to model developmental rates.
Field Measurements. Soil temperatures at a depth
of 2 cm, the approximate depth of grasshopper egg
pods in Alaska, were recorded at Þve locations southeast of Delta Jct., AK, in 2000, 2002, 2004, 2006, and
2008 (years of abundant grasshopper populations),
using thermocouples and data loggers (Onset Computing, Bourne, MA and Campbell ScientiÞc, Logan,
UT). The Þve locations were within a radius of 15 km
of one another and represented a range of herbaceous
plant communities common to roadsides, Conservation Reserve Program Þelds, and pastures known to
support abundant populations of grasshoppers. Common grasses (Cyperales: Poaceae) included native
species [Calamogrostis canadensis (Michx.) P. Beauv.,
Poa L. spp.] and introduced species (Bromus inermis
Leyss., Festuca rubra L., Hordeum jubatum L.). Common forbs included native perennial species [Chamerion angustifolium L. Holub (Myrtales: Onagraceae),
Cornus canadensis L. (Cornales: Cornaceae), Fragaria
virginiana Duchesne (Rosales: Rosaceae), and introduced annuals (Crepis tectorum L. (Asterales: Asteraceae), Taraxacum officialis F. H. Wigg. (Asterales:
Asteraceae)]. Woody plant species included Vaccinium L. spp. (Ericales: Ericaceae), Salix L. spp. (Salicales: Salicaceae), and Rosa acicularis Lindl. (Rosales:
Rosaceae). Thermocouples were installed in the
spring shortly after the ground thawed. Temperature
was recorded hourly throughout the growing season,
usually until mid-September.
Only M. borealis was abundant enough to provide data
for model veriÞcation. The postdiapause developmental
model was veriÞed by comparing predicted dates of egg
hatch to that observed in the Þeld. The prediapause
developmental model was veriÞed by comparing the
developmental stage predicted to be attained by embryos in the fall to that observed in eggs recovered from
natural populations in the Þeld. To estimate embryonic
development based on soil temperatures, rates of development were calculated for each hourly temperature
using either equation 1 or 2 (and divided by 24 to obtain
proportional development per hour). These estimates of
hourly proportional development were summed. Diapause initiation (in the case of prediapause development) or hatching (postdiapause development) was
predicted to occur on the date when cumulative
proportional development equaled 1.00. Postdiapause
development rate summation was initiated as soon as
the thermocouples were installed in the spring. Prediapause developmental rate summation was initiated on
1646
ENVIRONMENTAL ENTOMOLOGY
Table 1. Postdiapause development times (days to hatch ⴞ SE)
for grasshopper embryos at various temperatures
Species (population)
Temperature
M. borealis
(AK)
M. sanguinipes
(AK)
M. sanguinipes
(ID)
27/9a
27/12a
12
27/15a
15
27/18a
18
21
24
27
30
6.6 ⫾ 0.09
6.0 ⫾ 0.06
25.1 ⫾ 0.22
5.6 ⫾ 0.04
20.7 ⫾ 0.26
5.8 ⫾ 0.06
11.7 ⫾ 0.07
7.0 ⫾ 0.06
5.4 ⫾ 0.05
4.1 ⫾ 0.10
3.4 ⫾ 0.03
9.1 ⫾ 0.04
8.2 ⫾ 0.06
Ñb
6.8 ⫾ 0.04
26.5 ⫾ 0.22
7.1 ⫾ 0.09
13.7 ⫾ 0.07
9.1 ⫾ 0.06
6.4 ⫾ 0.06
4.8 ⫾ 0.04
3.5 ⫾ 0.04
9.9 ⫾ 0.07
10.7 ⫾ 0.07
Ñb
9.4 ⫾ 0.06
Ñb
9.1 ⫾ 0.09
18.7 ⫾ 0.08
11.4 ⫾ 0.10
7.9 ⫾ 0.05
5.8 ⫾ 0.05
4.7 ⫾ 0.04
a
b
Diurnally alternating temperatures, 12 h at each temperature.
No eggs hatched in this group.
the date when oviposition was estimated to have begun,
i.e., 7 d after 50% of natural populations of M. borealis had
reached the adult stage (determined from sweep samples). Because data are not currently available regarding
the duration of the preoviposition period, the 7-d interval
was based on preliminary data from the laboratory and
Þeld observations. Hatch date in Þeld was estimated from
the occurrence of Þrst instars in the Þeld. Sweep samples
were taken weekly at each of the Þve locations where soil
temperatures were recorded. Median hatch date was
estimated to be 3 d previous to the sample date with the
largest numbers of Þrst-instar M. borealis. Because populations of M. borealis were sparse at some of the locations in some years, data were pooled over all Þve locations, and a single hatch date was estimated for all
locations. The median predicted hatch dates from the
Þve locations were compared with the observed hatch
date.
The stage of development attained by prediapause
embryos in the Þeld was observed in 2004, 2006, and
2008. In the late fall, after soil temperatures had
dropped below 10⬚C consistently, soil samples were
sieved to recover egg pods laid during the summer.
Forty 0.25-m2 samples, 5 cm deep, were taken near the
locations where the soil temperature sensors were
installed. Eggs that were recovered were refrigerated
(5⬚C) until they could be examined to determine stage
of development. Because all eggs within a pod were at
a similar stage of development, each pod was considered a single observation.
Soil Temperatures and Climate Change Scenarios.
Average near-surface air temperatures during the
summer months in the interior of Alaska are estimated
Table 2.
Vol. 39, no. 5
to increase by 2.5Ð3.0⬚C before the end of the century
(Christensen et al. 2007). As an approximation of the
effects of increasing temperatures on the life cycle of
northern grasshoppers, three climate warming scenarios were simulated by adding 2, 3, or 4⬚C to each hourly
temperature recorded in the Þeld. Dates of completed
development were recalculated. To account for earlier hatching and more rapid nymphal development
under the climate change scenario, prediapause development rate summation was initiated 5, 7, and 9 d
earlier for soil temperature increases of 2, 3, and 4⬚C,
respectively. The degree of acceleration of nymphal
developmental rates were estimated using a model of
nymphal development for M. sanguinipes in Alaska
(Fielding 2004) using similar methods as for estimation of embryonic development described here.
Nymphal developmental times were not predicted to
be as strongly affected by increased air temperatures
compared with embryonic developmental times, because grasshoppersÕ internal temperatures are somewhat
independent of air temperatures because of thermoregulatory behavior. Finally, the latest date on which eggs
produced were predicted to hatch the following year and
the length of time available to produce offspring with a
1-yr life cycle were estimated.
Results
Stage of Diapause. Examination of embryos suggested that M. borealis diapauses in a slightly later
stage of development than M. sanguinipes. There
was no difference between the two populations of
M. sanguinipes in terms of femur length of the diapausing embryos (mean abdominal segment ⫾ SE,
6.8 ⫾ 0.05 and 6.7 ⫾ 0.05 for Alaska and Idaho,
respectively; F1,124 ⫽ 0.8, P ⫽ 0.36). Femurs of M.
borealis reached signiÞcantly further toward the
end of the abdomen than M. sanguinipes (7.4 ⫾ 0.05
versus 6.7 ⫾ 0.04 for M. borealis and M. sanguinipes,
respectively; F1,185 ⫽ 99.4, P ⬍ 0.0001).
Postdiapause Development. Eggs of M. borealis
hatched in the least amount of time at every temperature, and the Alaska population of M. sanguinipes
required less time than those from Idaho (Table 1).
Some of these differences could be attributed to diapausing at a later stage of development by M. borealis
but not for the two populations of M. sanguinipes.
Equation 1, without a developmental threshold, provided a slightly better Þt to the data for each group of
grasshoppers (Table 2; Fig. 1). For M. borealis, the log
likelihood was 29.0 and 27.7 for equations 1 and 2,
Maximum likelihood parameter estimates (SE) of temperature–postdiapause developmental rate functions
Species (population)
Parameter
␳
Tmax
⌬
Adjusted r2
M. borealis
M. sanguinipes
(AK)
M. sanguinipes
(ID)
0.147893 ⫾ 0.0764
41.3681 ⫾ 34.1825
6.74757 ⫾ 3.4798
0.917
0.186295 ⫾ 0.0616
38.8245 ⫾ 16.3112
5.364376 ⫾ 1.7713
0.957
0.237688 ⫾ 0.0444
34.1568 ⫾ 5.2365
4.20647 ⫾ 0.7844
0.959
October 2010
FIELDING AND DEFOLIART: EMBRYONIC DEVELOPMENTAL RATES IN GRASSHOPPERS
1647
Table 4. Median prediapause development times at various
temperatures (°C) measured as age in days (90% CI) at which
respiration rate drops below 0.30 (0.35 for M. borealis) ␮l/egg/h
Species (population)
Temperature
27/9a
27/12a
27/15a
27/18a
18
21
24
27
30
Fig. 1. Rate of postdiapause development as a function
of temperature. Squares, M. sanguinipes from Idaho; triangles,
M. sanguinipes from Alaska; circles, M. borealis. Open symbols, rates measuremented at constant temperatures; closed
symbols, rates estimated from alternating temperatures.
respectively. For M. sanguinipes from Alaska, the log
likelihood was 35.7 for equation 1, and the routine
failed to converge on parameter estimates for equation
2. For M. sanguinipes from Idaho, log likelihood was
35.7 and 34.5 for equations 1 and 2, respectively. Eggs
of M. borealis were the only group to hatch at 12⬚C
constant temperature. Both species from Alaska
hatched at 15⬚C constant temperature, but not M.
sanguinipes from Idaho (Table 1). Percentage hatch
was ⬎85% for all treatments where any hatching occurred within 45 d.
Predicted egg hatch dates for M. borealis differed
from observed egg hatch dates from 6 d too early to 5 d
too late but did not show any trend toward under- or
overestimation of hatch dates (Table 3). Field sampling was only conducted once per week, so the observed dates of hatch did not have Þne resolution, but
the 2 yr with earliest observed hatch dates (2002 and
2004) were also the two earliest dates predicted, and
the year with latest observed hatch (2008) was also the
latest predicted hatch date (Table 3). The model predictions of the magnitude of change resulting from
altered temperature regimens were consistent (Table
3) among years, providing a relative approximation of
the effects of altered temperature regimens on hatch
date. The model indicated that hatching could occur
3 or 4 d earlier with a 2⬚C increase in soil temperatures
a
M. borealis
M. sanguinipes
(AK)
M. sanguinipes
(ID)
44.6 (48Ð41)
44.3 (48Ð40)
35.9 (39Ð32)
35.0 (39Ð31)
77.2 (84Ð70)
47.6 (51Ð43)
28.4 (33Ð23)
22.3 (27Ð17)
19.5 (23Ð14)
44.5 (48Ð41)
41.9 (47Ð37)
35.2 (39Ð32)
34.9 (38Ð32)
73.8 (81Ð66)
46.9 (52Ð41)
29.2 (34Ð23)
24.1 (28Ð20)
16.9 (20Ð13)
35.9 (39Ð33)
34.8 (40Ð30)
29.7 (33Ð25)
28.1 (32Ð24)
62.4 (70Ð53)
35.8 (40Ð31)
21.6 (25Ð18)
17.8 (20Ð15)
12.9 (18Ð7)
Diurnally alternating temperatures, 12 h at each temperature.
and 5 or 6 d earlier with a 4⬚C increase (Table 3). If
the soil was 2⬚C cooler, hatching was predicted to
occur 3Ð 4 d later or 7Ð 8 d later if soils were 4⬚C cooler
(Table 3).
Prediapause Development. The relationship among
the different groups of grasshoppers in terms of speed
of prediapause development was reverse that of postdiapause development. Prediapause development,
from oviposition to diapause, was more rapid in the
Idaho population of M. sanguinipes compared with the
other grasshoppers (Table 4; Fig. 2) at temperatures of
21⬚C or greater. Prediapause development was very
slow at 18⬚C constant temperature for all populations.
Equation 2 (Table 5), with a deÞned developmental
threshold, provided a better description of developmental rates for M. borealis (log likelihood of 42.7 and
44.5 for equations 1 and 2, respectively). Equation 1
(Table 5) Þt the data better for M. sanguinipes from
Alaska (log likelihood of 42.5 and 34.5 for equations 1
and 2, respectively), whereas equation 2 Þt the data
better for the Idaho population of M. sanguinipes (log
likelihood of 39.8 and 42.6 for equations 1 and 2, respectively). The Alaska population of M. sanguinipes
Table 3. Observed median hatch dates and predicted hatch
dates of M. borealis based on soil temperatures and developmental
rates from laboratory
Year
Observed
median
hatch date
Predicted
median
hatch date
2000
2002
2004
2006
2008
4 June
28 May
29 May
1 June
5 June
31 May
30 May
23 May
3 June
10 June
Differences in predicted
hatch date (d) with
altered temp regimens
⫺4⬚C
⫺2⬚C
⫹2⬚C
⫹4⬚C
⫹8
⫹7
⫹8
⫹8
⫹8
⫹4
⫹3
⫹4
⫹4
⫹4
⫺4
⫺3
⫺3
⫺3
⫺3
⫺6
⫺5
⫺5
⫺6
⫺6
Fig. 2. Rate of prediapause development as a function of
temperature. Squares, M. sanguinipes from Idaho; triangles,
M. sanguinipes from Alaska; circles, M. borealis. Open symbols, rates measuremented at constant temperatures; closed
symbols, rates estimated from alternating temperatures.
1648
Table 5.
Parameter
␳
Tmax
⌬
␭
Adjusted r2
ENVIRONMENTAL ENTOMOLOGY
Vol. 39, no. 5
Maximum likelihood parameter estimates (SE) of temperature–prediapause developmental rate functions
Species (population)
M. borealis
M. sanguinipes (AK)
M. sanguinipes (ID)
0.195233 ⫾ 0.0771
34.90028 ⫾ 7.3576
5.12121 ⫾ 2.0218
0.00524 ⫾ 0.00686
0.985
0.153766 ⫾ 0.0404
47.2332 ⫾ 24.7931
6.501956 ⫾ 1.7060
Ñ
0.979
0.161437 ⫾ 0.0556
40.2742 ⫾ 9.0704
6.19182 ⫾ 2.1305
0.01003 ⫾ 0.00899
0.989
apparently did not reach zero developmental rate at
the temperatures tested (Fig. 2; Table 4), but developmental times for M. borealis at alternating temperatures of 27/9⬚C and 27/12⬚C, and for the Idaho population at 27/9⬚C, were approximately double those at
27⬚C constant, indicating that very little, if any development was taking place during the low temperature
phase of the diurnally alternating temperature treatments (Table 4).
Estimates of potential development based on soil
temperatures were consistent with observations of
embryonic development made in 2004, 2006, and 2008.
Embryos from pods recovered from the Þeld in 2004
spanned a wide range of development, from very early
stages to diapause stage (Fig. 3). Estimates of maximum potential development in 2004 ranged from 0.5
to 1.0 (diapause) over the Þve locations where soil
temperatures were monitored (Table 6). In 2006,
fewer pods were recovered (11 total), all of which
were in earlier stages of development (prekatatrepsis). Estimates of potential development in 2006 did
not exceed 0.52 across four locations where soil temperatures were measured (Table 6). In 2008, all eggs
were in a very early stage of development, as predicted
by the model (Table 6).
Over the 5 yr that soil temperatures were measured,
estimated maximum prediapause development ranged
from 0.24 to 1.08 (proportion of development to diapause) (Table 6). During the warmest year, 2004, two
locations had estimates of cumulative development
⬎1.0, suggesting the possibility that, given the current
climate regimen, in the warmest microsites during
Fig. 3. Relative proportion of prediapause embryonic
development completed at the end of the season in eggs of
M. borealis recovered from the soil in 2004, 2006, and 2008.
warmer years, a small portion of the population may
be univoltine. In the climate change scenario with 2⬚C
added to each hourly temperature and with development initiated 5 d earlier, only in the warmest year,
2004, was diapause stage predicted to be attained, but
with a correspondingly longer interval of time for
oviposition during which eggs may be expected to
hatch the following year. With a 3⬚C increase in soil
temperatures, diapause stage was predicted to be attained by some proportion of the population in 3 of the
5 yr. A 4⬚C increase in soil temperature was required
before diapause was predicted to be attained in most,
but not all, years (Table 6).
Discussion
The degree of prediapause development attained in
the Þrst year is the key process that determines voltinism in high-latitude populations of grasshoppers.
Because grasshopper eggs are immobile, they cannot
engage in behavioral thermoregulation and are passive
subjects of the thermal environment. Therefore, it is
relatively straightforward to predict the effects of reduced temperatures caused by shading of the soil
surface or increased temperature caused by climate
change on embryonic developmental time. The results
indicated that, with increasing soil temperatures, univoltinism is likely to become more common in grasshopper populations that are currently primarily semivoltine. Comparison of the 5 yr of soil temperatures
also show considerable variation among years in the
amount of prediapause development attained by
grasshopper embryos. Annual variation in temperatures suggests that changes in the life cycle of these
insects may be erratic, which adds to the complexities
of evaluating population responses to altered temperature regimens. The effects of lower soil temperatures
on prediapause development were not considered
here, because changes of only a few degrees would not
slow down prediapause development enough to extend their life cycle an additional year.
In contrast to the wide variation in the degree of
prediapause development in the fall, observed hatching dates were fairly consistent among years, probably
because all embryos are in the same, late stage of
development in the year of their hatching. This information is valuable not so much for predicting egg
hatch in any given yearÑ egg hatch tends to be very
consistent in this region and a calendar may be as
useful a tool as any for predicting time of hatchÑ but
rather as a means of predicting life cycles.
October 2010
Table 6.
FIELDING AND DEFOLIART: EMBRYONIC DEVELOPMENTAL RATES IN GRASSHOPPERS
1649
Estimated maximum potential proportion of prediapause development completed at the end of the season by M. borealis
Year
Start datea
Actual soil temperature
⫹2⬚C
⫹3⬚C
⫹4⬚C
2000
2002
2004
2006
2008
23-Jul
16-Jul
5-Jul
21-Jul
20-Jul
0.10Ð0.32
0.32Ð0.59
0.41Ð1.08 (3 d)
0.21Ð0.53
0.10Ð0.24
0.24Ð0.55
0.59Ð0.95
0.68Ð1.50 (17 d)
0.38Ð0.90
0.25Ð0.49
0.39Ð0.74
0.81Ð1.22 (6 d)
0.91Ð1.83 (24 d)
0.53Ð1.15 (6 d)
0.42Ð0.72
0.53Ð0.95
1.08Ð1.52 (12 d)
1.19Ð2.15 (30 d)
0.69Ð1.42 (11 d)
0.60Ð1.00 (1 d)
Development was predicted using actual soil temperatures recorded at Þve locations and with 2, 3, or 4⬚C added to each hourly temperature
to simulate climate change scenarios. Data show predicted prediapause development from the coolest and warmest of the Þve locations in each
of 5 yr. A 1-yr life cycle is predicted by values ⱖ1.0. Values in parentheses are the length of time during which eggs produced are predicted
to hatch the following spring.
a
Estimated date of Þrst oviposition by adult grasshoppers. This date was advanced by 5, 7, and 9 d for the 2, 3, and 4⬚C climate warming
scenarios, respectively.
The models estimated a delay of 8 d in hatch date
resulting from a 4⬚C decrease in soil temperatures. As
with prediapause development, predicting the effects
of temperature on hatching date is straightforward,
but understanding just how such a delay would affect
grasshoppersÕ Þtness is not known. In Saskatchewan,
Pickford (1960) found that M. sanguinipes that
hatched earlier had greater nymphal survival and
greater adult fecundity. Although no explanation for
the effects of hatching date on nymphal survival was
given, Pickford (1960) suggested that cooler temperatures late in the season accounted for reduced fecundity in those that hatched later. In western Canada, hatching by M. sanguinipes may be spread out over
a period of several weeks, and the experiments of
Pickford (1960) encompassed a similar range of hatch
dates. In Alaska, hatching dates are much more compressed. Although it may seem that less variabililty in
time of hatch would reduce the amount of variability
in Þtness costs because of hatching phenology, it is also
likely that, because of the shorter warm season, the
Þtness costs may increase at a steeper rate with delayed hatching.
The model without a developmental zero (equation
1) provided the better Þt in each case of postdiapause
development (Table 2). Eggs of M. sanguinipes did not
hatch within 45 d at low constant (12⬚C) temperatures, but rates estimated from the ßuctuating regimens indicate that some development was proceeding. Fisher (1994), working with a population of M.
sanguinipes from Montana, reported hatching at 12⬚C
after 65 d, but only a low proportion hatched (⬍20%).
When using only constant temperature treatments, it
might be assumed, incorrectly, that if no hatching
occurs, no development is taking place. This may be
one reason for the widespread use of models that use
a linear, degree-day model that assumes a developmental zero (Gage et al. 1976, Kemp and Sanchez
1987). This observation of a higher temperature
threshold for eclosion than for general development
may also partially explain the sudden pulse of hatching
that typiÞes grasshopper populations in Alaska. Cool
soil temperatures may allow embryos to develop to the
point of hatching, but actual eclosion may be delayed
until soil temperatures exceed the threshold for
eclosion.
In predicting the effects of altered temperature
regimens, we assumed that grasshopper embryos are
passive entities subject to the thermal environment.
This paradigm ignores the question of whether embryonic developmental rates are Þxed or may evolve
in response to changing temperature regimens. The
comparison of prediapause and postdiapause developmental rates between the two populations of M.
sanguinipes suggests that embryonic developmental
rates may be heritable and that genetic diversity for
this trait exists. The population from Idaho had the
most rapid prediapause development but was the
slowest to hatch postdiapause. This pattern could be
partially explained by diapausing at an earlier stage of
development, but no external differences in diapause
stage were observed between the two populations of
M. sanguinipes. When comparing total embryonic developmental times (sum of prediapause and postdiapause developmental times), the Idaho population
completed development in less time than the Alaskan
population at every temperature examined (Tables 1
and 4). If there is a selective advantage to an annual
life cycle over a 2-yr life cycle, it is conceivable that
faster embryonic developmental rates could evolve in
northern populations of grasshoppers, hastening the
switch to an annual life cycle. Studies of life histories
of insects suggest that changes in numbers of generations per year can be abrupt (Powell and Logan 2005,
Fielding 2006).
The results presented here indicate that predicted
changes in the climate may be enough to make a
univoltine life cycle in Melanopline grasshoppers in
the interior of Alaska more common. This study focused solely on the effect of temperature on rates of
embryonic development. Although voltinism in these
insects is determined almost solely by abiotic conditions, the net effect on population dynamics of a
change from primarily a biennial life cycle to univoltism is a much more complex question. Logic and empirical research suggest that abiotic and biotic factors
will interact to determine the effects of a changing
climate on herbivorous insect populations. For example, ambient temperatures may inßuence the intensity
of intraspeciÞc competition (Laws and Belovsky
2010), alter predatorÐprey relations and subsequent
trophic cascades (Chase 1996), and disrupt phenological synchrony of herbivorous insects and their host
plants (van Asch et al. 2007). Hence, a more complete
understanding of the response of high-latitude grasshopper population dynamics to altered thermal
1650
ENVIRONMENTAL ENTOMOLOGY
conditions will depend on further investigations into
factors regulating populations in this environment.
Acknowledgments
S. Gneiting and W. Connor assisted with Þeld and laboratory work. The manuscript was improved by the comments
and suggestions of A. Pantoja, D. Branson, and Þve anonymous reviewers.
References Cited
Alexander, G. 1973. Uniformity of successive populations of
an alpine grasshopper with a two-year diapause. Entomol.
News 84: 321Ð328.
Altermatt, F. 2010. Climatic warming increases voltinism in
European butterßies and moths. Proc. R. Soc. B Biol. Sci.
277: 1281Ð1287.
Branson, D. H., A. Joern, and G. A. Sword. 2006. Sustainable
management of insect herbivores in grassland ecosystems: new perspectives in grasshopper control. Bioscience 56: 743Ð755.
Chase, J. M. 1996. Abiotic controls of trophic cascades in a
simple grassland food chain. Oikos 77: 495Ð506.
Cherrill, A., and M. Begon. 1991. Oviposition date and pattern of embryogenesis in the grasshopper Chorthippus
brunneus (Orthoptera, Acrididae). Holarctic Ecol. 14:
225Ð233.
Christensen, J. H., B. Hewitson, A. Busuioc, A. Chen, X. Gao,
I. Held, R. Jones, R. K. Kolli, W.-T. Dwon, R. Laprise, et
al. 2007. Regional climate projections, pp. 847Ð940. In S.
Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K. B.
Averyt, M. Tignor, and H. L. Miller (eds.), Climate
change 2007: the physical science basis. Contribution of
Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge, United Kingdom.
Church, N. S., and R. W. Salt. 1952. Some effects of temperature on development and diapause in eggs of Melanoplus bivittatus (Say) (Orthoptera: Acrididae). Can. J.
Zool. 30: 173Ð184.
Fielding, D. J. 2004. Developmental time of Melanoplus sanguinipes (Orthoptera: Acrididae) at high latitudes. Environ. Entomol. 33: 1513Ð1522.
Fielding, D. J. 2006. Optimal diapause strategies of a grasshopper, Melanoplus sanguinipes. (http://www.insectscience.
org/6.02).
Fielding, D. J. 2008. Diapause traits of Melanoplus sanguinipes and M. borealis. Ann Entomol. Soc. Am. 101: 439 Ð
448.
Fisher, J. R. 1994. Temperature effect on postdiapause development and survival of embryos of three species of
Melanoplus (Orthoptera: Acrididae). Ann. Entomol. Soc.
Am. 87: 604 Ð 608.
Gage, S. H., M. K. Mukerji, and R. L. Randell. 1976. A predictive model for seasonal occurrence of three grasshopper species in Saskatchewan (Orthoptera: Acrididae).
Can. Entomol. 108: 245Ð253.
Gray, D. R., J. A. Logan, F. W. Ravlin, and J. A. Carlson. 1991.
Toward a model of gypsy-moth egg phenologyÑ using
respiration rates of individual eggs to determine temperature time requirements of prediapause development.
Environ. Entomol. 20: 1645Ð1652.
Heliovaara, K., and R. Vaisanen. 1984. The biogeographical
mystery of the alternate-year populations of Aradus cinnamomeus (Heteroptera: Aradidae). J. Biogeogr. 11: 491Ð
499.
Vol. 39, no. 5
Kemp, W. P., and N. E. Sanchez. 1987. Differences in postdiapause thermal requirements for eggs of two rangeland
grasshoppers. Can. Entomol. 119: 653Ð 661.
Kreasky, J. B. 1960. Extended diapause in eggs of high-altitude species of grasshoppers, and a note on food-plant
preferences of Melanoplus bruneri. Ann. Entomol. Soc.
Am. 53: 436 Ð 438.
Kukal, O., and P. G. Kevan. 1987. The inßuence of parasitism on the life history of a high arctic insect, Gynaephora
groenlandica (Wocke) (Lepidoptera: Lymantriidae).
Can. J. Zool. 65: 156 Ð163.
Lactin, D. J., N. L. Holliday, D. L. Johnson, and R. Craigen.
1995. Improved rate model of temperature-dependent
development by arthropods. Environ. Entomol. 24: 68 Ð75.
Laws, A. N., and G. E. Belovsky. 2010. How will species
respond to climate change? Examining the effects of
temperature and population density on an herbivorous
insect. Environ. Entomol. 39: 312Ð319.
Logan, J. A., D. J. Wollkind, S. C. Hoyt, and L. K. Tanigoshi.
1976. An analytic model for description of temperature
dependent rate phenomena in arthropods. Environ. Entomol. 5: 1133Ð1140.
Martin-Vertedor, D., J. J. Ferrero-Garcia, and L. M. TorresVila. 2010. Global warming affects phenology and voltinism of Lobesia botrana in Spain. Agric. Forest Entomol.
12: 169 Ð176.
Mikkola, K. 1976. Alternate-year ßight of northern Xestia
species (Lep., Noctuidae) and its adaptive signiÞcance.
Ann. Entomol. Fenn. 42: 191Ð199.
Moore, H. W. 1948. Variations in fall embryological development in three grasshopper species. Can. Entomol. 80:
83Ð 88.
Morimoto, N., O. Imura, and T. Kiura. 1998. Potential effects of global warming on the occurrence of Japanese
pest insects. Appl. Entomol. Zool. 33: 147Ð155.
Mukerji, M. K., S. H. Gage, and R. L. Randell. 1977. Inßuence of embryonic development and heat on population
trend of three grasshopper species in Saskatchewan (Orthoptera: Acrididae). Can. Entomol. 109: 229 Ð236.
Musolin, D. L. 2007. Insects in a warmer world: ecological,
physiological and life-history responses of true bugs
(Heteroptera) to climate change. Global Change Biol. 13:
1565Ð1585.
O’Neill, K. M., B. E. Olson, M. G. Rolston, R. Wallander, D. P.
Larson, and C. E. Seibert. 2003. Effects of livestock grazing on rangeland grasshopper (Orthoptera: Acrididae)
abundance. Agric. Ecosys. Environ. 97: 51Ð 64.
Pickford, R. 1960. Survival, fecundity and population
growth of Melanoplus bilituratus (Wlk.) (Orthoptera:
Acrididae) in relation to date of hatching. Can. Entomol.
92: 1Ð10.
Pierson, F. B., and J. R. Wight. 1991. Variability of nearsurface soil temperature on sagebrush rangeland. J. Range
Manage. 44: 491Ð 497.
Powell, J. A., and J. A. Logan. 2005. Insect seasonality: circle
map analysis of temperature-driven life cycles. Theor.
Popul. Biol. 67: 161Ð179.
Salt, R. W. 1949. A key to the embryological development of
Melanoplus bivittatus (Say), M. mexicanus mexicanus
(Sauss.) and M. packardii Scudder. Can. J. Res. D 27:
233Ð235.
Salzen, E. A. 1960. The growth of the locust embryo. J.
Embryol. Exp. Morph. 8: 139 Ð162.
SAS Institute, 2003. SAS/STAT userÕs guide. Release 9.1.
SAS Institute, Cary, NC.
Slifer, E. H. 1932. Insect development IV. External morphology of grasshopper embryos of known age with a
known temperature history. J. Morphol. 53: 1Ð22.
October 2010
FIELDING AND DEFOLIART: EMBRYONIC DEVELOPMENTAL RATES IN GRASSHOPPERS
Tobin, P. C., S. Nagarkatti, G. Loeb, and M. C. Saunders.
2008. Historical and projected interactions between climate change and insect voltinism in a multivoltine species. Global Change Biol. 14: 951Ð957.
van Asch, M., P. H. van Tienderen, L.J.M. Holleman, and
M. E. Visser. 2007. Predicting adaptation of phenology
1651
in response to climate change, an insect herbivore example. Global Change Biol. 13: 1596 Ð1604.
Zar, J. H. 1999. Biostatistical analysis, 4th ed. Prentice-Hall,
Englewood Cliffs, NJ.
Received 30 November 2009; accepted 26 May 2010.