Competition and growth in Aedes aegypti larvae: effects of distributing food inputs over time

Male and female mosquito larvae compete for different subsets of the yeast food resource in laboratory microcosms. Males compete more intensely with males and females with females. The amount and timing of food inputs alters both growth and competition, but the effects are different between sexes. Increased density increases competition among males. Among females, density operates primarily by changing the food/larva or total food; this affects competition in some interactions and growth in others. Food added earlier in the life span contributes more to mass than the same quantity added later. After a period of starvation larvae appear to use some of the subsequent food input to rebuild physiological reserves in addition to building mass. The timing of pupation is affected by the independent factors and competition, but not in the same way for the two sexes, and not in the same way as mass at pupation for the two sexes. There is an effect of density on the timing of pupation for females independent of competition or changes in food/larva or total food. Male and female larvae have different larval life history strategies. Males grow quickly to a minimum size, then pupate, depending on the amount of food available. Males that do not grow quickly enough may delay pupation further to grow larger, resulting in a bimodal distribution of sizes and ages. Males appear to have a maximum size determined by the early food level. Females grow faster than males and grow larger than males on the same food inputs. Females affect the growth and competition among males by manipulating the number of particles in the microcosm through changes in feeding behavior. Mosquito larvae appear to have evolved to survive periods of starvation and take advantage of intermittent inputs of food into containers.

changing the food/larva or total food; this affects competition in some interactions and growth in 23 others. Food added earlier in the life span contributes more to mass than the same quantity added 24 later. After a period of starvation larvae appear to use some of the subsequent food input to rebuild 25 physiological reserves in addition to building mass. The timing of pupation is affected by the 26 independent factors and competition, but not in the same way for the two sexes, and not in the same 27 way as mass at pupation for the two sexes. There is an effect of density on the timing of pupation for 28 females independent of competition or changes in food/larva or total food. Male and female larvae 29 have different larval life history strategies. Males grow quickly to a minimum size, then pupate, 30 depending on the amount of food available. Males that do not grow quickly enough may delay pupation 31 further to grow larger, resulting in a bimodal distribution of sizes and ages. Males appear to have a 32 maximum size determined by the early food level. Females grow faster than males and grow larger than 33 males on the same food inputs. Females affect the growth and competition among males by 34 manipulating the number of particles in the microcosm through changes in feeding behavior. Mosquito

269
The aliquot treatment and the interaction between aliquots and timespan have significant R squared 270 values, but below most of the food, density and timespan contrasts; the interesting 3-way interaction 271 between food, density and aliquot has the second lowest R squared of the significant contrasts.

272
273 Table 1 shows the three significant interaction contrasts involving food and density ( The standardized discriminant function coefficients do not lend themselves to easy interpretation of the 306 biology underlying the differences; however, that there are differences across the dependent variables 307 suggests that males and females respond differently to the independent factors, and that the Prime and 308 Average individuals respond differently within sexes. The MANOVA and these coefficients describe the 309 relationships among the dependent variables and will be reconsidered again along with the individual 310 ANOVA results later.  Table 2 shows the three significant interaction contrasts for the three attributes of the food supply: food 318 level, aliquot, and timespan. The 3-way interaction between these three factors (FxAxT) is not 319 significant in the MANOVA, but the three 2-way interactions are significant. These interactions do not 320 include density, so do not describe competition. Instead, they describe the growth of larvae on different 321 food regimens: the two food levels, and the distribution of the food inputs in time (aliquot and 322 timespan). The FxA contrast has large positive coefficients for the Prime male mass and age at pupation 323 and smaller coefficients for Survival and the Prime female mass, and for the Prime female age at 324 pupation (negative). The coefficient for Survival is the largest for that variable across the MANOVA. The 325 R squared value for this contrast is only slightly larger than that of the FxDxA contrast. The FxA 326 interaction represents the residual after the FxDxA interaction is accounted for, so this is the effect of 327 food level and aliquot on the growth of the larvae after the competition has been explained. The FxT 328 interaction has a high coefficient for the Prime male age at pupation followed by smaller coefficients for 329 Survival, Prime male mass at pupation, and Average male mass at pupation. The coefficients for the 330 three variables measuring female mass and age are smaller still. The R squared value for this contrast interaction is accounted for, so this is the effect of food level and timespan on the growth of larvae after 333 the competition has been explained. The AxT interaction has large positive coefficients for Average 334 male mass and Average female mass and a smaller positive coefficient with Prime male age at pupation 335 and Survival and a smaller negative coefficient with Prime female age at pupation. The R squared value 336 is the fifth ranked across the interactions (and the three lower ranked values are much smaller). The

337
AxT interaction represents the residual after the FxAxT and DxAxT interactions are accounted for (see 338 below for a description of the DxAxT interaction). It describes the effect of aliquot and timespan, the 339 distribution of food in time, on the growth of larvae. These three interactions have the three highest 340 coefficients for Survival across the interaction contrasts; the parameters of the food supply affect 341 Survival more than any interaction with density affects it. The Prime male age at pupation is also 342 affected by these three interactions. The Prime male mass and to a lesser extent, the Prime female 343 mass, are affected by the FxT and FxA interactions, but not by the AxT interaction; however, Average 344 male mass and Average female mass are affected by FxT and AxT, but not by FxA. Prime female age at 345 pupation is slightly affected by FxA and AxT, but not by FxT. As in Table 1, the discriminant function 346 coefficients suggest that male and female larvae grow differently in response to the independent 347 factors. The MANOVA and these coefficients describe the relationships among the dependent variables 348 and will be reconsidered again along with the individual ANOVA results later.  Table 3 shows the two significant interaction contrasts for density and the non-food level attributes of 356 the food supply (DxAxT and DxT). The 2-way interaction, DxA, is not significant in the MANOVA. The

357
AxT interaction does not include density and is described in Table 2. Density has an interaction with 358 aliquot and timespan, and with timespan, that is independent of the food x density ( relationship between food availability and density, such that: a) growth of larvae is good at high 372 availability (earlier delivery: 4 aliquots and/or 3 day timespan) and low density; b) growth of larvae is 373 less good at high availability and high density, or at low availability (later delivery: 2 aliquots and/or 6 374 day timespan) and low density; and c) growth of larvae is much worse at low availability and high 375 density. As in Tables 1 and 2, the discriminant function coefficients suggest that male and female larvae the relationships among the dependent variables and will be reconsidered again along with the The experiment looks for interactions between the aliquot and timespan treatments and food x density 388 (representing different competitive regimes). The main effects are all significant as expected and 389 Survival is not strongly affected by any treatments or interactions, as hoped. There is no 4-way 390 interaction, so the aliquot treatment and the timespan treatment do not interact jointly with food x 391 density. There is only one 3-way interaction with a high R squared value: food x density x timespan. The 392 2-way interactions across these 3 factors are also significant with high R squared values: food x density, 393 food x timespan, and density x timespan. These four interactions account for the highest R squared 394 interactions across the MANOVA. The 3-way interaction between food x density x aliquot is significant, 395 but has a low R squared value. There is a small interaction between the aliquot treatment and 396 competition.

398
The food x aliquot interaction, the food x timespan interaction, and the aliquot x timespan interaction 399 describe the growth of larvae rather than competition among them. These interactions are not related 400 to competition because they do not include density as a factor, but both aliquot and timespan affect the 401 rate of delivery of food to the larvae and thus the growth and eventual pupation of the individuals.

403
The 3-way interaction, DxAxT, has the lowest R squared value across the significant contrasts.

774
Individual females grow as fast as possible in response to environmental conditions. As 1 st instars, food 775 may be unlimited, but chance and the initial size of the larva (or egg size, or some other uncontrolled 776 factor in the experiment) determine how fast the larva grows. At some point, this larva molts to the 2 nd 777 instar. The larva may delay molting in order to grow larger because food is available, but there is also an 778 advantage to molting in order to use the larger feeding apparatus of the 2 nd instar. This trade off 779 between growing larger and molting should also be true for the 3 rd and 4 th instar molts. Prime males minimize the time to pupation rather than maximizing size at pupation. Prime male age at 798 pupation is affected by food and competition. In two treatments, Prime males pupate slightly after day 799 5 apparently because food is abundant (3 day timespan); both of these pupate at larger masses than 800 their 6 day timespan peers. In the two most competition treatments, Prime males pupate slightly after 801 day 5, with the 6 day timespan treatment pupating mostly on day 6, but still before the addition of food 802 at the end of day 6. The mass at pupation of Prime males is determined by food, but can be modified by 803 delaying pupation to take advantage of abundant food, or to continue growing because food is scarce.

804
The aliquot treatment does not interact with competition for Prime males; there is no effect of dividing 805 the total food into 2 or 4 aliquots. In the FxD interaction, the mass of the Prime male is determined by 806 other factors, but the age at pupation is affected by the residual competition after the higher order 807 interactions have been removed. Prime male age at pupation is earliest with the high food, high density 808 treatment, a little later in the least competition and low food, low density treatments and 809 disproportionately delayed in the most competition treatment. and in the 6 day timespan treatments, they receive additional food and appear to grow larger as a resulting in larger Prime males and smaller non-Prime males compared to the equivalent food/larva 818 treatment, low food, low density, 3 day timespan. Despite the same food/larva (4 mg food/larva) the 819 Prime males are smaller and the Average male mass is larger at the low density than at the high density.

820
There is also a residual effect of competition on the size of non-Prime males (in the FxD interaction).

821
Non-Prime males decrease in size in the treatments with increased competition, but they are 822 disproportionately small in the most competition treatment (FxD). Despite the small size of the non-

823
Prime males in the most competition treatment (FxD), at least one of them grows larger than the Prime 824 male so that the Average male mass is larger than the Prime male mass. This is similar to the 825 observations in the FxDxT interaction. There is no timespan treatment here, so the increase in size of 826 the non-Prime male over the Prime males must be due to the release from competition rather than 827 additional food. There is also evidence that competition among males is more intense in the high food, 828 high density treatment than in the low food, low density treatment (with the same food/larva). The

829
Prime male is larger, the Average male mass is smaller and the difference between them is larger in the 830 high food, high density test tubes compared to the low food, low density test tubes. due to the large input on day 3 similar to that observed at the high food level.

913
At the low food level and the 6 day timespan, the Prime females do not have enough food to pupate 914 until after the last aliquot on day 6. They grow larger and faster on the 4 aliquot treatment than the 2 915 aliquot treatment. Despite equal food and a longer larval period, they do not grow as large as the Prime 916 females in the 3 day timespan treatments. Food early in the larval period is more important than later 917 additions for both size and growth rate.

919
The Average female masses are also arranged in 3 groups with large gaps between each group. In the 920 group with the highest food level the Average female mass in the 4 aliquot, 3 day timespan is the largest 921 and the closest to that of the Prime female mass. In this treatment with abundant food, the non-Prime 922 females do better than in any other treatment. There is a bigger difference between the Average 923 female mass in this treatment and the next largest Average female, than between the corresponding 924 Prime females, so the non-Prime females do better on the 4 aliquot treatment than on the 2 aliquot 925 treatment (at the high food level and 3 day timespan).

927
There is a larger gap in size between the 3 largest Average females and the middle group of Average 928 females (compared to the same gap for the Prime females). This is due to the smaller size of the 929 Average females in the high food, 2 aliquot, 6 day timespan relative to the Prime female in that Large initial inputs of food allow the Prime female to grow faster and to dominate the food supply.

951
Smaller, regular inputs (4 aliquot treatment) allow the non-Prime females to grow larger than equivalent 952 amounts of food in fewer, larger inputs. The food level is the most important factor for the Prime 953 females, followed by the interaction between aliquot and timespan. The interaction between aliquot 954 and timespan is relatively more important for the non-Prime females, especially at the highest and 955 lowest food levels. The final aliquot in the 6 day timespan appears essential to Prime females at the 956 lowest food level, but does not appear to benefit the Prime females in the other treatments (because 957 they pupate before the addition of the food), or the non-Prime females (perhaps because the size 958 distribution is fixed by the molt into the fourth instar).
this examines the effect of food and aliquot on growth not competition. This interaction is significant 962 for the Prime male mass and age at pupation and the Average male mass at pupation in the ANOVAs.

963
The three variables: Prime male mass and age, and Average male mass, are not significant in the 964 FxDxAxT, FxDxA, or FxAxT interactions described earlier. This is the highest order interaction between 965 food and aliquot for males.

967
This interaction shows that the growth of males is affected by the both the amount of food and the 968 number of aliquots it is divided into, separately from competition for food (food x density) and 969 timespan. Both Prime and non-Prime males grow better on 4 aliquots than 2 aliquots. It appears to be 970 the food delivered in the day 2 to day 4 period (the third aliquot) that causes all the male larvae to grow 971 larger in the 4 aliquot treatment (the main effect of aliquot). This interaction between food and aliquot 972 arises because males grow larger than expected and take longer in the low food, 4 aliquot treatment.

973
Males are growing rapidly and filtering particles, so the added particles accelerate their growth, and this 974 has a larger effect at low food levels than at high food levels. Prime males defer pupation and extend 975 their growth in response to the third aliquot of food, and grow larger than Prime males at the same food 976 level and the 2 aliquot treatment. Although the Prime and Average males are smaller in the low food 977 treatment, the third aliquot has a larger impact on growth at the low food level compared to the high 978 food level.

979
The low food treatment causes the females to switch from active filtering to retaining particles in their 980 guts, reducing the available particles. This reduces the size and size distribution among the females, and 981 also reduces the size and the size distribution among males. Males are smaller than projected at the low 982 food level, especially in the 2 aliquot treatment, but the size distribution of the males is also Average male masses in the low food, 4 aliquot treatment are larger than in the low food, 2 aliquot 986 treatment, and are almost as large as their projected values. This is due to the third aliquot of food that 987 is delivered in the middle of the larval growth period (day 2 or day 4, depending on timespan 988 treatment). Prime and non-Prime males grow larger but the relative advantage of the Prime female 989 over the Prime male is smaller, so the Prime male benefits more from this third aliquot food addition 990 than the non-Prime males. Males appear to actively filter at all food levels, unlike females, which 991 appear to switch from active filtering to retaining particles in their guts. In this situation, the males 992 respond sooner to the third aliquot and the Prime male grows larger relative to the Prime female. The

993
Prime male also grows larger relative to the non-Prime males due to exponential growth processes and 994 the initial size distribution before the third aliquot. The Prime males extend their larval period to grow 995 larger on the extra food, and pupate latest in this treatment (across this interaction).

997
The Average male masses are in the same order as the Prime male masses, so all males appear to 998 benefit from the third aliquot, but the Average male mass in the low food, 2 aliquot treatment is the 999 same as the Prime male mass. The non-Prime males in this treatment grow as large as the Prime males.

1000
They don't grow as large as the non-Prime males in the low food, 4 aliquot treatment; the third aliquot 1001 delivers more food earlier in the larval lifespan. The non-Prime males in the low food, 2 aliquot 1002 treatment must take advantage of the additional food in the final aliquot (on day 3 or day 6) to grow as 1003 large as the Prime male. This suggests that there is a minimum mass at pupation for males that is 1004 determined by environmental conditions before day 3. Male larvae grow until they reach that 1005 minimum, or grow larger than the minimum, even delaying pupation, in response to greater availability 1006 of food.

1008
The main effect of the 4 aliquot treatment increases the size of all larvae compared to the 2 aliquot 1009 treatment, but for males at the low food level, it provides a disproportionate benefit. between the rates for females and those for males is at the highest total food. The differences decrease 1090 as the size of the pupae and the total food decrease, suggesting that this is due to exponential growth 1091 processes.

1092
Males and females respond to food differently after the effects of competition and aliquot have been 1093 removed by the higher order interactions. Females grow linearly in response to the total food after day 1094 4. They required 16 mg of food in order to pupate and delay pupation until after day 6 in the low food, 1095 6 day timespan treatment. Males also grow larger in response to more total food, and they delay 1096 pupation both when food is available and when it is limiting, but pupate mostly on day 5. Males in the 1097 low food, 6 day timespan treatment are much smaller than the other males, but the non-Prime males 1098 grow larger than the Prime males on the final addition of food on day 6. This appears to be due to 1099 exponential growth processes in response to the total food, but modified by different pupation triggers 1100 between the two sexes.

1139
Prime males pupate before the day 6 food input, so there is even less food for them than for the Prime 1140 and Average females.

1141
In contrast to the Prime females, the Prime males pupate latest in the 4 aliquot, 6 day timespan 1142 treatment, but still before the final input of food on day 6. The Prime males in both 6 day timespan 1143 treatments delay their pupation to grow larger, but the ones with more food (3 aliquots instead of 1 1144 aliquot) delay longer and grow larger than the ones in the 2 aliquot treatment. The estimated growth 1145 rates follow the Prime male mass; this interaction affects the mass 5 times more than the Prime male 1146 age.

1147
The difference between the Prime male mass and the Average male mass indicates changes in the size 1148 distribution of males and the relative advantage of the Prime male over the non-Prime males; it is 1149 almost entirely affected by the timespan. The difference between the Prime and Average males is larger test tubes with the 3 day timespan, so this is possibly due to exponential growth processes, another 1152 contrast to the observation for females.

1154
All the Prime males pupate before the addition of the final aliquot on the 6th day. In the 6 day timespan 1155 treatments some of the non-Prime males grow larger on the final aliquot and the mass of the Average 1156 males is more similar to that of the Prime males. This is likely due to increased food for the non-Prime

1219
DxA is not significant in the MANOVA, but is significant for Prime female age at pupation in the ANOVA.

1298
Because this interaction involves density and timespan, an attribute of the food supply, it is possible that 1299 competition is involved. Food level (total food) and density are independent factors in the experiment, 1300 and they jointly affect competition among the mosquito larvae, but growth and competition among the 1301 larvae are also affected by food/larva, which is necessarily confounded with both food level and density.

1302
The DxT interaction reflects the residual effects of density and timespan on growth and competition.

1303
Timespan clearly affects the amount of food in the test tubes on a daily basis, so could be affecting the 1304 competition among larvae as well as the growth of larvae. the Prime male to delay pupation when food is available; neither food/larva nor total food consistently 1380 predict this response. Prime males consistently delay pupation when competition is intense, but do not 1381 delay it long enough to benefit from the additional food on day 6 in the day 6 timespan treatments.

1383
Males versus females. Density and timespan jointly affect the mass and age variables for males and 1384 females, but they affect them differently for the two sexes. Females grow largest, fastest, the Prime 1385 females pupate earliest, and the relative advantage of the Prime female over the non-Prime females is 1386 smallest at the low density and 3 day timespan. Either the 6 day timespan or the high density reduce 1387 the size, the growth rate, and increase the age at pupation and the relative advantage of the Prime relative advantage of the Prime female over the non-Prime females. The females do disproportionately 1391 well in the low density, 3 day timespan and disproportionately poorly in the high density, 6 day 1392 timespan treatments. The interaction between density and an attribute of the food supply could be an 1393 indication of competition independent of the food level (FxD), but there is no evidence for it. As the 1394 food availability early in the larval period changes due to a longer timespan and higher density, the 1395 females all pupate at smaller sizes, grow more slowly, and the size distribution increases. The simple 1396 explanation is that these are the result of exponential growth processes rather than competition, 1397 because females should switch from active filtering to retention at low food levels (and they appear to 1398 do so in other interactions in this experiment), and that would cause the size distribution to compress 1399 rather than to increase as observed.

1401
Males are also largest and grow fastest at the low density, 3 day timespan treatment, and smallest, grow 1402 slowest and pupate latest at the high density, 6 day timespan treatment. They differ from the females 1403 in that the earliest age at pupation is in the low density, 6 day timespan treatment and relative 1404 advantage of the Prime male over the non-Prime males is more affected by timespan than by density.

1405
The size distribution of the males increases with density in the 3 day timespan treatments, similar to 1406 that of the females. However, the size distribution of males decreases with increasing density in the 6 1407 day timespan treatments, probably due to the added food after the Prime male pupates. In contrast 1408 with the females in this interaction, the Prime male mass and age, the Average male mass, and the 1409 difference between the Prime and Average male masses jointly resemble the pattern in the set of FxD 1410 interactions that describe competition (above). This suggests that males are competing for food in 1411 response to the changes in availability caused by the timespan treatment independently of the food Females grow larger and faster than males in all treatments. The relative difference in size of the Prime 1416 female over the Prime male and that of the Average female over the Average male could also indicate 1417 competitive interactions. These differences correspond to the relative advantage of the Prime female 1418 over the Prime male and of the non-Prime females over the non-Prime males. With one exception, the 1419 advantage decreases as the mass of the males and females decreases, suggesting exponential growth 1420 processes. The Prime male in the high density, 3 day timespan treatment is relatively larger than in the 1421 high density, 6 day timespan (so the difference between the Prime female and the Prime male is 1422 smaller). Prime males in this treatment are also relatively larger than the Average males. The high 1423 density with the 3 day timespan increases the relative advantage of the Prime male over the non-Prime 1424 males, but also as compared to the females. This interaction does not affect competition among 1425 females, but it appears to affect competition among males and possibly between males and females. to measure the mass and age at pupation in response to density (1-3 larvae/vial) and food level (2 mg -1445 5 mg food/larva). This analysis was not possible because of the differential mortality shown in Table 5.

1446
The raw data are presented in S2 Dataset.
1447 Survival at the lowest density (1 larva per vial) was lower than in the other two treatments, and favored 1450 the survival of males over females (8 males, 2 females). The comparison between the pupae raised 1451 alone and those raised with competitors was not possible because too many cells were empty. No 1452 statistical tests were performed on the survival data (the data did not conform to the original 1453 experiment due to the differential mortality). as for the vials with 2 and 3 larvae. Males at the lowest density take longer to pupate, while females 1456 take less time to pupate at the lowest density. No statistical tests were performed on the mass and age 1457 data (the data did not conform to the original experiment due to the differential mortality). assigned to a test tube, so there is some variability due to that. Some larvae died before the second 1477 input of food, others died after the second input of food during the experiment, and another group was 1478 moribund after 15 days (the end of the experiment). (See Tables 8 and 9 below.) Table 7 shows the 50 1479 males and 40 females that were analyzed in the experiment (56% male). In addition to those, 10 males 1480 pupated before the second input of food and were excluded from the analysis. Overall, there were 60 1481 males and 40 females (60% male) that survived to pupation including those males that pupated before 1482 the second food input. 1502 Table 8 shows the number of deaths by treatment and by time period. 50 of 150 larvae died (33%).

1503
There were not enough deaths for tests of significance across the time period or food input treatments.

1504
The 6 or 8 day period of starvation is expected to be the largest source of differential mortality. 23 1505 larvae died in the 6 day treatment and 27 larvae died in the 8 day treatment across the entire 1506 experiment.  Table 9 shows the age at death by treatment. In addition to the 31 larvae that died during the 1510 experiment, another 19 were moribund after 15 days and never pupated. 9 of the 31 larvae in Table 8 1511 died before the second food input. No analyses were performed on this data.  to the r squared value in the ANOVA (although the R squared values are not additive, unlike the r second largest R squared value across the interactions is the 2-way interaction between food 1 x delay; 1528 however, there is a significant 3-way interaction that includes these two factors: food 1 x delay x sex.

1529
The 3-way interactions mean that the amount of food added and the length of the delay (period of 1530 starvation) jointly affect the growth of the larvae, and this effect is different depending on the sex of the 1531 larva. The discriminant function coefficients indicate the weight of each of the two dependent variables 1532 on the contribution to the significance of that contrast. In most of the contrasts the mass variable is 1533 more important than the age variable. In those contrasts where this is not true, the ANOVA (see Table   1534 11) shows that the mass variable is not significantly affected by the contrast, but the age variable is 1535 significantly affected. (These contrasts are: the main effect, food 2; the interaction food 1 x sex; the 1536 interaction food 2 x sex). to the second food input jointly affect the mass and age at pupation, and this effect is different for the the MANOVA, and there is also a significant 3-way interaction for these factors (food 1 x delay x sex). squared values than the food 1 x delay for age. These are: food1 x sex and food 2 x sex, and they are 1552 only significant for age, not for mass. Males and females respond differently to the food input and delay 1553 treatments; both mass and age at pupation are affected, but the treatments affect them differently.
1554 Third experiment result summary 1559 The detailed analyses of the interactions for the dependent variables, mass and age at pupation, across 1560 the MANOVA and ANOVAs are presented in S3 Text and accompanying tables, S45 Table-S60 Table,  Food 2 x delay x sex. Sex is more important to mass at pupation than food input or delay for this 1564 interaction. Food is more important to mass at pupation than delay. Females grow larger on the day 6 1565 delay than on the day 8 delay at both food levels. Males grow larger on the day 8 delay at the lower 1566 food inputs, but grow to the same size at the 3 mg food input. This suggests that males reach a size 1567 determined by environmental conditions and pupate as soon as they reach that size, while females grow 1568 as large as possible. Sex is more important to age at pupation than either food input or delay for this 1569 interaction. Delay is more important to age at pupation than food input amount.

1570
Females pupate earliest in the test tubes with the 3 mg food input and the 6 day delay. The females in 1571 the other 3 treatment combinations pupate almost a day later. The similarity of the ages at pupation of 1572 the females in the treatments with lower food inputs, and in the 3 mg food input with the longer delay 1573 suggests that females may have an optimal maximum duration of the larval period. They grow to 1574 different sizes depending on the food input and the delay, but they take almost the same amount of 1575 time after the food input to do so.

1576
Males also pupate earliest in the test tubes with the 3 mg food input and the 6 day delay, but the males 1577 in the test tubes with the lower food input and the 6 day delay pupate shortly afterwards, followed by 1578 the lower food input and the 8 day delay, then the 3 mg food input and the 8 day delay. Males 1579 postpone pupation at the 3 mg food input and the day 8 delay to grow to the same size as the males in The estimated growth rates of females are higher than those of males in both 3 mg food input 1583 treatments, but lower than those of males in both lower food input treatments. Males and females 1584 grow more slowly on the day 8 delay than the day 6 delay at both food input amounts. The additional 2 1585 days of starvation in the day 8 delay treatments results in the lower growth rates; males take longer to 1586 reach the same size (at 3 mg food input) while females grow for longer, but still pupate at a smaller 1587 mass than at the 6 day delay. This suggests that both males and females are improving their depleted 1588 physiological state at the same time as they are adding mass in the day 8 delay treatments.

1589
Food 1 x delay x sex. Food input amount (1 mg vs 2 mg) appears to be more important to the mass at 1590 pupation than sex and also more important than the delay treatment; in the prior interaction, food 1591 input amount (1 mg + 2 mg vs 3 mg) is less important than either sex or delay. Food is probably limiting 1592 the growth of larvae in the test tubes with the 1 mg second food input, and may be limiting the growth 1593 of larvae in the test tubes with the 2 mg second food input.

1594
The delay treatment has little effect on the mass of either males or females at the 1 mg food input. The 1595 delay treatment has a larger effect at the 2 mg food input, and the effect on males and females is 1596 opposite. Males pupate at the largest mass with the 2 mg food input and the day 8 delay, while females 1597 pupate at the largest mass with the 2 mg food input and the day 6 delay. The males in the 2 mg food 1598 input, day 8 delay test tubes pupate at the same size as the males in the 3 mg food input test tubes.

1599
They apparently reach the maximum size as determined by the environment. Since these males receive 1600 a different second food input than the 3 mg food input males, the maximum size of males must have 1601 been set before the second food input, so before day 6. Because the males in the 2 mg food input, day 8 1602 delay treatment grow as large as males in both 3 mg food input treatments, males are probably not 1603 limited by food at the 2 mg food input. Females grow larger on the 3 mg second food input than on the 1605 The estimated growth rate indicates that males and females respond differently to the effects of 1606 starvation (delay) and the amount of food in the second input. The difference in growth rate across the 1607 two delay treatments for both sexes is due to the age at pupation (not significant in this interaction).

1608
The growth rates support the idea that the larvae are rebuilding physiological reserves that do not show 1609 up as mass in the treatments with the 2 mg food input and the day 8 delay.

1619
The effect of the delay also changes as the food input increases. At the 1 mg second food input, there is 1620 little effect of delay on the pupal mass of females. At the 2 mg and 3 mg food inputs, the day 8 delay 1621 treatment females are 0.24 mg to 0.25 mg smaller than the day 6 treatment females.

1622
There is also an effect of delay and food on age at pupation. The day 6 delay treatment females pupate 1623 earlier as the size of the second food input increases. The incremental food reduces the age at pupation 1624 as it increases the mass at pupation, and the effect of the 1 mg of food on both mass and age is less with 1625 each additional increment. In contrast, the age at pupation for the day 8 delay treatment females The additional two days of starvation change the way that females respond to additional food. They still 1628 grow larger on the extra food, but they don't grow as large and they seem to grow for an amount of 1629 time that is not affected by the size of the food input. Females may be replenishing physiological 1630 resources that were used during the extra two days of starvation (day 8 delay), or they may have some 1631 maximum larval period after the food input, or both.

1632
Food 1 x delay. This interaction shows the residual effect of the size of the second food input and the 1633 delay on the growth of larva regardless of sex. The amount of the second food input is more important 1634 to the mass at pupation than the delay in this interaction. The mass at the 1 mg food input is much 1635 smaller than expected while the mass at the 2 mg food input is only a little larger than expected, and the 1636 difference due to the delay is larger at the 2 mg food input. The second food input is more important 1637 than the delay for age at pupation as well as mass at pupation; the larvae pupate earlier on the 2 mg 1638 food input than the 1 mg input, and earlier with the day 6 delay than the day 8 delay. The second food 1639 input increases the mass and decreases the age at pupation, but the delay treatment increases both the 1640 mass and the age.

1641
Food 2 x delay. This interaction between food 2 and delay compares the two lower second food inputs

1642
(1 mg and 2 mg) with the highest second food input (3 mg) crossed with the two delay treatments. The 1643 size of the second food input is also more important than the delay in this interaction. The lower food 1644 inputs grow larger on the day 8 delay (see above) but the 3 mg food inputs grow larger on the day 6 1645 delay. The age at pupation is not significantly affected by this interaction; the larvae in the 3 mg food 1646 input and day 6 delay pupate much earlier than the other treatments, while those in the 3 mg food input 1647 and day 8 delay pupate latest. results in larger pupae, but each additional increment increases the size of the pupae by a smaller 1651 amount. The age at pupation is only significant for the food 1 contrast, the two lower food inputs. The 1652 age at pupation drops much more with the 2 mg second food input in the day 6 delay treatment, so the 1653 effect of the additional food on age at pupation changes depending on the period of starvation (delay 1654 treatment). The age at pupation decreases again in the 3 mg second food input with the day 6 delay 1655 treatment, but increases slightly in the 3 mg food input with the day 8 delay.

1656
The larvae appear to be using the food for something other than added mass in the day 8 delay 1657 treatments. It is likely that the early growth and molts of the larvae determine the largest size of the 1658 larvae, but that is similar for both the day 6 delay treatment and the day 8 delay treatment in this 1659 experiment. This suggests that the larvae are using some of the food and extra feeding time to improve 1660 their physiological condition rather than building mass in the day 8 delay treatments. Perhaps the larvae 1661 (both sexes) use the first milligram of food primarily for growth in mass, and the second milligram allows 1662 them to grow larger and faster, but the third milligram allows them to replenish the physiological 1663 resources that were depleted during the extra two days of starvation.

1664
Food 1 x sex and food 2 x sex. These interactions are significant for age at pupation. Males pupate 1665 earlier than females at all three food inputs. Males pupate earlier than expected and females pupate 1666 later than expected (from the main effects). The age at pupation for males differs from the expected 1667 value by a similar amount for all three food inputs. The age at pupation for females differs from the 1668 expected value by a smaller amount as the food input increases. Males and females deviate from the 1669 expected values in different ways.

1670
Males grow faster than females except at the 3 mg food input (estimated growth rate). The males in the grow at the same rate as the males in the 2 mg food input. Males do not appear to be food limited in 1673 either the 2 mg or the 3 mg second food inputs.

1674
Females grow faster, larger and pupate earlier with each increment of food. They grow more slowly 1675 than males at the 1 mg food input, at almost the same rate at the 2 mg food input, but faster than males 1676 at the 3 mg food input. This could be due to the apparent maximum size of males (2.27 mg in this 1677 experiment) or to exponential growth processes (the larger size of females) or both. Females appear to 1678 grow in response to the amount of food in the second food input; they may still be food limited at the 3 1679 mg food input whereas the males appear to grow equally fast in both the 2 mg and the 3 mg food input, 1680 suggesting that even the 2 mg food input is more than sufficient food for them.

1681
Food, starvation and sex. Male and female mosquito larvae respond to food inputs after periods of 1682 starvation in different ways. Males appear to grow to a size determined before day 6 of the larval 1683 period in response to the addition of food at the end of a period of starvation. The size of the food 1684 input, and the period of starvation (delay) affect both the mass and age at pupation. At the shorter 1685 period of starvation (the day 6 delay treatment), males grow larger with each increment of food and 1686 pupate earlier. At the longer period of starvation, males take longer to pupate than at the shorter 1687 period of starvation and the relationship between pupal mass and incremental food is obscured because 1688 even at the 2 mg food input, males grow to the 2.27 mg maximum size for this experiment. Males do 1689 not appear to be food limited at either the 2 mg or 3 mg food inputs.

1690
Females appear to grow larger in response to the amount of the second food input. Each additional 1691 increment of food contributes less to the mass at pupation, and the longer period of starvation further 1692 reduces the contribution of each increment to the pupal mass. The age at pupation of females at pupation. For the longer period of starvation (the day 8 delay) females grow larger on each 1696 increment of added food, but not as large as in the comparable shorter period treatment, and they take 1697 longer to grow that large. Females may have a maximum larval period that causes them to pupate, 1698 resulting in smaller pupae despite equivalent amounts of food. Females may be food limited at the 2 mg 1699 food input and perhaps even at the 3 mg food input.

1700
The larvae may be using the food inputs differently in the shorter and longer starvation treatments.

1701
Both sexes may be rebuilding physiological reserves that were depleted during the additional 2 days of 1702 starvation. Male and female larvae raised in isolation respond differently to identical conditions. Each 1703 increment of added food is less valuable than the previous increment as measured by mass at pupation.

1704
The value of food increments differs across sexes. Males and females alter the age at pupation 1705 differently in response to both food input amount and the period of starvation. Males appear to have a 1706 maximum size at pupation that is determined before day 6. Females appear to have a maximum 1707 duration of larval period that is affected by the timing of the second food input. Male and female larvae 1708 may delay pupation to rebuild physiological reserves at higher levels of the second food input. This 1709 delay appears to differ across sexes.

1711
These three experiments provide insight into how the environment affects 4 aspects of mosquito larval 1712 life: competition among mosquito larvae, the survival and growth of the larvae, and the triggers that 1713 initiate pupation. Competition, growth and the pupation triggers differ across the sexes; this suggests 1714 that survival may also differ across the sexes, but because the larvae that died were not identified by

Survival across the three experiments
1718 The levels of the independent factors were selected so that survival would be high. Survival is high in 1719 the first and the third experiments. In the second experiment, the much lower survival at the lowest 1720 density (1 larva per vial) prevents the analysis of the data and suggests that larvae facilitate the feeding 1721 of their peers. Also in the second experiment, the second lowest density (2 larvae per vial) appears to 1722 have reduced survival, but the experiment was not designed to test the effect of density on survival, so 1723 no statistical analyses were performed.

1724
Competition in the first experiment has a small effect on survival (FxDxT accounts for 5% of the The aliquot treatment in the first experiment also affects survival; larvae survive better on 2 aliquots 1731 than on 4 aliquots (12% of the variance). This is opposite of the effect of the aliquot treatment on mass 1732 and age at pupation (larvae grow better on 4 aliquots than on 2 aliquots). There are no interactions 1733 between aliquot and the other treatments for survival; the effect of aliquot is independent of the 1734 interactions described previously. The 4 aliquot treatment provides more food early in the larval period 1735 than the 2 aliquot treatment, so this result is symmetrical to the observations above that too much food 1736 too early in the larval period reduces survival, but the effect of the number of aliquots is independent of 1737 food, density, competition and timespan.
Overall, the survival observations suggest that mosquito larvae are adapted to low levels of food and 1739 conditions of starvation, and that these experiments provide food levels that may be at the high end of 1740 those encountered in nature.

1741
Competition among females 1742 The first experiment revealed that competition among females affects both mass and age at pupation, There is also a residual effect on the Prime female age at pupation; this increases disproportionately 1784 with increasing competition after the higher order interactions have been removed. The mass at 1785 pupation of the Prime female is determined by the higher order interactions, and by the main effects, 1786 but the age at pupation is increased by increased levels of competition in this residual interaction.

1787
Competition among males 1788 The first experiment reveals that competition among males affects both mass and age at pupation, and 1789 is affected by timespan. In the most significant interaction describing competition, FxDxT, Prime males

1795
The most intense competition between Prime and non-Prime males appears to be in the high food, high 1796 density, 3 day timespan treatment where the difference between the Prime and Average male mass is 1797 greatest, and the Prime male has the largest numerical advantage over the non-Prime males. In the 1798 most competition treatment, both Prime and Average males are the smallest, but the Average male is 1799 larger than the Prime male, indicating that the non-Prime males grow larger after the Prime male 1800 pupates and on the additional food in the last food input at the end of day 6. The effect of this 1801 interaction on the non-Prime males suggests that food/larva, competition among females, release from 1802 competition after the pupation of the Prime male, and additional food on day 6 all contribute to the 1803 outcome of competition as measured in the mass at pupation. Competition among males differs from The Prime male age at pupation is not in the same order as either the total food in the test tube, the 1806 food/larva, or the Average male mass. 7 of the 8 treatments pupate early (between 5.0 and 5.12 days), 1807 but the Prime males in the most competition, 6 day timespan treatment pupate later, after 5.70 days. In 1808 two cases, Prime males with greater access to food delayed pupation slightly perhaps to increase in size 1809 or improve physiological status. For males, age and mass at pupation are affected differently by 1810 competition and the factors: food, density and timespan. Males pupate earlier than females, and at a 1811 smaller size, and extend their larval life in response to too little food (most competition), and also when 1812 food is relatively abundant.

1813
The FxDxA interaction has no effect on competition among males, in contrast to females. After the 1814 effects of the FxDxT interaction are removed, there is a residual effect of competition on the Prime male 1815 age and the Average male mass. Because the Prime male mass is not significantly affected by this 1816 interaction, the residual effect is on the non-Prime males. These are disproportionately smaller in the 1817 most competition treatment than in the other treatments, and they are also disproportionately reduced 1818 in size relative to the Prime male in the high food, high density intermediate competition treatment. In 1819 the most competition treatment, the Average male mass is again larger than the Prime male mass. In 1820 this case, there is no timespan treatment, so the difference in size reflects only the release from 1821 competition, not the addition of the final input of food on day 6. These are the same effects as seen in 1822 the higher order interaction, FxDxT, but there is a residual effect on only the non-Prime males. As with 1823 female larvae, non-Prime males are more affected by competition than Prime males.

1824
There is also a residual effect on the Prime male age at pupation; this increases disproportionately with 1825 increasing competition after the higher order interactions have been removed. The mass at pupation of There are two other interactions that appear to affect competition among males, DxAxT and DxT. The

1829
Average male mass appears to be affected by the DxAxT interaction in the same way that it is affected 1830 by the FxDxT interaction. In this case, the size of the Average males corresponds to the estimated 1831 growth rate of the Prime male (the Prime male age and mass are not significantly affected by this 1832 interaction, but the growth rate is an indication of the competitive stress on the Prime male). The

1833
Average male grows larger than the Prime male in the treatment with the most competition (lowest 1834 growth rate). The non-Prime males grow larger than the Prime male due to the release from 1835 competition after the Prime male pupates, and on the additional food after the final input on day 6. The

1861
The Prime male delays pupation in the two treatments with the most total food, and the treatment with 1862 the least food/larva and most competition. This is consistent with observations in other interactions, 1863 especially FxDxT.

1864
Competition between males and females 1865 The first experiment shows that males compete differently from females and are affected differently by 1866 the factors, aliquot and timespan. Females actively filter particles at high food levels and switch to 1867 retention at lower food levels, changing the nature of competition among females. Males appear to 1868 filter particles without switching to retention at low food levels, so they are disproportionately affected 1869 by the particle level when females switch to retention. Males compete more intensely among 1870 themselves and may be competing for a subset of the particles, perhaps because of their smaller size.

1871
When competition is most intense among female larvae, the Prime male does better relative to the 1872 Prime female, but the non-Prime males do worse. Non-Prime males experience a release from timespan treatments. The effect of the extra food is many times larger than the effect of the release 1875 from competition (0.14 mg for food plus release in FxDxT versus 0.01 mg for just the release in FxD).

1876
The age at pupation is affected differently by competition than the mass at pupation in both sexes, and 1877 the effect on age at pupation is different between the sexes. The Prime female age follows the total 1878 food in the test tube while the Prime male age at pupation does not. The Prime male age is earlier, 1879 more simultaneous, and the small deviations appear to be due to food scarcity or food abundance.

1880
There is a residual effect of competition on age at pupation for both sexes after the higher order

1893
The Prime females in this treatment already experience an unlimited food supply, so more food does 1894 not increase their masses. This is probably true for the Prime males as well; both Prime male mass and food and extra food should increase their mass, but there is no effect of this interaction on Average 1898 male mass. This suggests three possibilities: the size of the non-Prime males is already determined 1899 before day 3 and the additional food does not affect it; the additional food is monopolized by the non-1900 Prime females so there is none for the non-Prime males; there is already so much food for the non-

1901
Prime males as well as the Prime male and Prime female that the additional food has no effect on any of 1902 them. Both Prime and non-Prime males grow larger in response to additional food after day 3, so the 1903 suggestion that the size of males is determined before day 3 is not supported by evidence in other   1904 interactions. Other interactions suggest that increased food benefits all larvae, so the third possibility 1905 seems likeliest.

1906
In the FxDxA interaction only the females experience competition. Both Prime and Average female 1907 masses follow the food/larva with the total food affecting the mass when the food/larva is equal across 1908 treatments. The 4 aliquot treatment results in more food by day 4 than the 2 aliquot treatments and 1909 this leads to larger females and smaller differences between Prime and Average females in the 4 aliquot 1910 treatments. The additional food due to the third aliquot reduces competition among females, but has 1911 no effect on the competition among males or on the competition between males and females. This 1912 suggests three possibilities similar to the ones in the FxDxAxT interaction above: the size of both Prime 1913 and non-Prime males is determined before day 3 and the additional food has no effect on it; the 1914 additional food is monopolized by the females so there is none for the males; there is already so much 1915 food for the males that the additional food has no effect on them. Both Prime and non-Prime males 1916 grow larger in response to additional food after day 3, so the suggestion that the size of males is females. The Prime female mass is also significantly affected by this interaction, but the effect appears 1933 to be on growth rather than competition among females. This interaction further suggests that males 1934 and females are competing for different subsets of the food resource.

1935
In the DxT interaction, the Prime male and the Average male compete similarly to the FxD interaction.

1936
The timespan treatment affects the food environment independently of the food level or number of 1937 aliquots and changes the competitive environment among males, but not among females. At the high 1938 density, the 6 day timespan has a greater effect on competition among males, reducing the size of Prime 1939 and Average males and delaying the age at pupation. Females appear to be affected by the food/larva 1940 (mass variables) and the total food (age at pupation), but these effects appear to be due to exponential 1941 growth processes rather than competition. This suggests that males and females compete differently 1942 for different subsets of the food resource and also that males respond to the timing of food inputs and they extend their larval life until there is sufficient total food (16 mg for 8 larvae) or food/larva (2 mg food/larva) for them to pupate. Males actively filter at all food/larva levels and Prime males pupate 1946 later at low food levels, but still before the final input on day 6. Non-Prime males grow larger than 1947 Prime males on the release from competition and on the additional food on day 6, but the increment in 1948 this residual interaction is on the order of that in the FxD interaction (the release from competition).

1949
The mechanisms of competition are the same as in previous interactions, but this suggests that males 1950 are more responsive to inputs than females, perhaps because it takes longer for females to switch from 1951 retention to actively filtering, or perhaps because they require a larger number of particles to initiate 1952 the switch. (DxAxT, and DxT) and the third one, DxA, is only significant for Prime female age at pupation.

1987
The second experiment also investigated competition and growth, but due to differential mortality no 1988 statistical analyses were made.

2093
The DxA interaction is only significant for Prime female age at pupation. The age at pupation is inversely 2094 related to the food/larva after day 4, but the Prime female pupates disproportionately early in the low 2095 density, 2 aliquot treatment. Since timespan is not a factor in this interaction, the 3 day and 6 day 2096 timespans are equivalent. In contrast to interactions including timespan, the difference across the 2097 treatments appears to be the amount of food in the initial input. Prime females in the low density, 2 2098 aliquot treatment have more food/larva on day 0 than the other treatments; this accelerates their 2099 growth and results in the early pupation relative to the expected values.

2100
The DxT interaction represents the residual effect on Prime female mass and age and Average female 2101 mass after the higher order interaction effects are removed. These interactions include competition 2102 (FxDxT, FxDxAxT) and growth (DxAxT). The residual effect on male larvae appears to be competition, 2103 but the residual effect on females appears to affect growth. There is an almost linear correspondence 2104 between the food/larva after day 4 and the masses at pupation for both Prime and Average females.

2105
Despite the low food/larva levels in some treatments, the ongoing food inputs over the timespan appear 2106 to allow the females to actively filter particles. The incremental increases in food across these four 2107 treatments are not regular and the corresponding increases in Prime and Average female mass are also The Prime female age at pupation in the DxT interaction follows the total food after day 4 rather than 2110 the food/larva after the effect of the FxDxT interaction is removed. This is similar to the age at pupation 2111 in the FxDxT interaction (competition), but the DxT interaction appears to describe exponential growth 2112 processes for the female mass variables. For the Prime female, age at pupation may be more affected 2113 by the food environment than by competition.

2114
In the third experiment, individual larvae receive 1 mg of food on day 0 and then 1 mg, 2 mg, or 3 mg of 2115 food on day 6 or on day 8. Females do not pupate on less than 2 mg of food (total), which is consistent 2116 with the 2 mg food/larva necessary in the first experiment (16 mg total food for 8 larvae

2182
In the FxA interaction, Prime males grow larger in the 4 aliquot treatment than in the 2 aliquot 2183 treatment and aliquot has a larger effect at the low food level than at the high food level. It appears 2184 that the third aliquot (day 2 to day 4) accelerates the growth of the Prime male. Although both Prime 2185 and Average male masses are smaller at the low food level, the Prime male in the 4 aliquot treatment is 2186 relatively large compared to both the Average male mass and the Prime female mass. This suggests that 2187 males feed on a different subset of the food resource, probably based on particle size, and that the third 2188 aliquot enables the Prime male to grow at the expense of the non-Prime males. This is the result of 2189 exponential growth processes rather than competition. The Average male masses are in the same order 2190 as the Prime male masses, so all males appear to benefit from the 4 aliquot treatment. The Average 2191 male mass in the low food, 2 aliquot treatment is the smallest across this interaction, but it is the same 2192 as the Prime male mass, so the non-Prime males grow as large as the Prime males (probably on the food 2193 added on day 6 in some of the test tubes). There is no release from competition in this interaction, but 2194 there is additional food in some test tubes. Despite the late input of additional food, the non-Prime 2195 males don't grow as large as the non-Prime males in the low food, 4 aliquot treatment (equivalent food, 2196 delivered earlier in the larval life). This suggests that there is a minimum mass at pupation for males 2197 that is determined by environmental conditions around day 3. Male larvae grow until they reach that 2198 minimum, or grow larger than the minimum, even delaying pupation, in response to greater availability early food inputs. Additional food late in the larval life allows growth, but that growth is capped 2202 because of physiological or physical limitations (size of the feeding apparatus or other body parts).

2203
Prime males delay their age at pupation in the 4 aliquot treatment relative to the 2 aliquot treatment at 2204 the same food level. Prime males take longer to pupate at the lower food level, and delay pupation in 2205 the 4 aliquot treatment disproportionately longer at the lower food level.

2206
In the FxT interaction, the residual effect on female growth is an almost linear relationship between the 2207 total food after day 4 and the pupal mass of females. Prime male mass and Average male mass are also 2208 directly related to the total food in three of the four treatments (0.2 mg increments versus 0.5 mg 2209 increments for the females). The Prime male at the lowest total food is disproportionately small, but 2210 still pupates before the final input on day 6. The Average male mass in this treatment is larger than the 2211 Prime male mass, so the non-Prime males grow larger than the Prime male on the final input of food on 2212 day 6. Competition has been removed from this interaction (by FxDxT) so this represents only the effect 2213 of additional food on day 6. The reason behind the disproportionately small size for the males in the 2214 low food, 6 day timespan treatment is probably the retention of food particles by the females reducing 2215 the apparent food level for the males. The Prime male pupates before the addition of food on day 6; 2216 the non-Prime males and females grow larger on the day 6 food input, but do not grow as large as larvae 2217 that received the same amount of food earlier.

2218
Prime males delay pupation at the highest total food and at the two lower total food levels. They grow 2219 for longer and reach a larger size when food is most available, and also delay pupation when food is less 2220 available, but they still pupate on day 5 or day 6.

2221
In the AxT interaction, the residual effect on females indicates that more food early in the larval life 2222 increases the mass for both Prime and Average females. There are no higher order interactions for the grow disproportionately largest in the 2 aliquot, 3 day timespan treatment and disproportionately 2225 smallest in the 2 aliquot, 6 day timespan treatment. The two 4 aliquot treatments are intermediate in are affected differently than for females. At the shorter delay, equivalent to the 6 day timespan 2249 treatment in the first experiment, males grow larger and pupate earlier as the second food input 2250 increases. At the longer delay (8 days, 2 extra days of starvation), males grow larger than at the shorter 2251 delay, but take longer to pupate. Males at the longer delay and 2 mg second food input, and at both 2252 delays and the 3 mg second food input, pupate at 2.27 mg, which appears to be the maximum size of 2253 males in this experiment. Because this maximum does not seem dependent on the size of the second 2254 food input, it must be determined by environmental conditions before day 6.

2255
Summary of the growth of male larvae 2256 The mass and age at pupation of males are more affected by competition than by growth; all 2257 interactions involving density appear to describe competition among males. FxT is the residual after 2258 competition has been removed; male mass appears to be directly related to the total food in the 2259 treatments with the exception of lowest total food. In the 3 treatments with the higher total food, each 2260 increment of food adds 0.2 mg to the mass of the Prime male and the Average male; this is only 40% of 2261 the increase that females realize, suggesting that males are using a subset of the food that females 2262 experience. The males in the low food, 6 day timespan treatment are disproportionately smaller than 2263 the males and females in other treatments and the females in this treatment. The non-Prime males in 2264 this treatment also grow larger than the Prime males on the additional food added on day 6. At the 2265 lowest total food, the feeding behavior of the female larvae constrains the growth of the male larvae 2266 independently of competition.

2267
Males grow largest on 4 aliquots in the FxA interaction and the effect is more pronounced at the low 2268 food level. The extra food in the third aliquot appears to contribute to the growth of Prime and non-Both mass and age at pupation may be affected by the physiological state of the male larvae; this is 2294 needed to explain the differences observed at different food inputs after the two periods of starvation 2295 (in the third experiment).

2296
Differences in the growth of males and females 2297 Food is the most important factor in the growth of mosquito larvae of both sexes, and the first 2298 experiment shows that the amount of food during the late larval period (3 rd and 4 th instars, days 2-4) is 2299 the most important determinant of the mass at pupation. Multiple food inputs during the early larval 2300 period benefit all larvae to some extent. Large inputs on day 0, and on day 3 can affect the ultimate 2301 mass at pupation and the size distribution of pupae by altering the feeding behavior of females (actively 2302 filtering versus retention), which affects the availability of particles to both female and male larvae.

2303
The interactions describing the growth of female larvae partially overlap those describing the growth of

2307
FxAxT, DxAxT and AxT show the effect of aliquot and timespan on the growth of females: an input of 2308 food early in the larval period contributes more to growth than the same amount of food later; inputs 2309 during the late instars accelerate growth; more, smaller inputs favor the growth of all female larvae, 2310 while fewer inputs alter the size distribution of females depending on the timing of the inputs. Aliquot 2311 and timespan affect the total food and food/larva in the test tubes; total food and food/larva affect the 2312 feeding behavior of females and this changes the availability of particles in the test tubes. This is the The aliquot treatment does not affect competition among Prime males, but it does affect the growth of 2338 Prime and Average males. The FxA interaction indicates that males grow disproportionately large at the 2339 low food level and the 4 aliquot treatment compared to the 2 aliquot treatment. Independent of 2340 density and timespan, the 4 aliquot treatment delivers more food early in the larval life than the 2 2341 aliquot treatment. The difference is likely the third aliquot of food on day 2 or day 4. This is similar to 2342 the effect of the third aliquot on the growth of females in the FxAxT and AxT interactions. This suggests 2343 that the early food inputs are more important to the growth of males than later ones. Furthermore, the 2344 Prime male mass is disproportionately large at the low food, 4 aliquot treatment compared to both the 2345 Average male mass and the Prime female mass. The comparison against the Average male mass can be 2346 explained by exponential growth processes, but the comparison against the Prime female suggests that 2347 males compete for a different subset of the food supply than the females.

2348
Some non-Prime males in the FxA interaction do grow larger on the day 6 food input and the Average 2349 male mass is the same as the Prime male mass at the low food level, 2 aliquot treatment. This suggests 2350 that there is a minimum mass at pupation for males that is determined by environmental conditions 2351 around day 4 (the day the third aliquot is added for half the timespan treatments). Despite receiving 2352 much more food on day 6, the non-Prime males do not grow as large as the non-Prime males in the 4 2353 aliquot treatment at the same food level. Males grow until they reach this minimum mass, then grow 2354 larger if there is abundant food, delaying pupation in order to grow. Prime males pupate before day 6.

2355
Some non-Prime males grow larger on the additional food after day 6 (in some treatments), but the food 2356 added on day 6 is not as beneficial as food added earlier.

2357
The amount of food added after a period of starvation affects the mass at pupation of females. The size growth trajectories of each instar, the data from these experiments is consistent with female larvae larger than the Prime male in some circumstances when food is added to the test tubes on day 6 (after In the absence of sufficient food to complete pupation, both males and females wait for additional food 2453 inputs. Both males and females pupate at a larger size and earlier as the amount of the late food input 2454 increases, but the length of the period of starvation affects the value of the food and the contribution to 2455 pupal mass. Males grow larger and take longer to pupate after the longer period of starvation 2456 compared to the shorter (6 day) period. Females do not grow as large after the longer period of 2457 starvation and appear to be limited to a specific interval of growth after the late food input. Males 2458 reach their maximum size and pupate, while females grow until they reach their maximum larval period 2459 and then pupate.

2460
Both males and females appear to use some of the second food input to replenish internal physiological 2461 factors after a period of starvation; the pupal mass does not reflect the entire value of the late food 2462 input and this is more apparent after the longer period of starvation. This means that the triggering of 2463 pupation is dependent on at least: age or time, mass, environmental food conditions, and internal 2464 physiological status. Furthermore, the maximum size of pupae is probably also constrained by early life 2465 history due to the size of the head capsule, the feeding apparatus or some other physical or 2466 physiological feature determined by the molt to the 4 th instar.

2467
Mosquito larvae could trigger larval molts and pupation based on interactions with other larvae in 2468 addition to the factors mentioned above. The mass at pupation of the Prime male is clearly affected by 2469 the sex ratio in the microcosm (see experiment 2 in [1]), which suggests the possibility of direct 2470 interactions between larvae. Females are not affected by the sex ratio in that experiment; this suggests 2471 that female larvae are not interacting with other larvae (male or female) and that they pupate only in 2472 response to their internal state and the environmental factors. This is supported by the lack of any were the case, the same pattern of distribution of sizes should appear. There is no evidence that these 2547 three interactions, DxAxT, DxA, or DxT, affect competition among females. They appear to describe the 2548 growth of females.

2549
Competition among males also results in a pattern of size distributions. Prime males respond to total 2550 food and food/larva similarly to Prime females, but the Average males do not correspond to the Average 2551 females. Males appear to be competing for a different subset of the yeast particles than females; this is 2552 probably related to the different sizes of male and female larvae and their associated feeding structures.

2553
Males compete more intensely with other males and females with other females. Females do affect 2554 competition among males by reducing the low particle numbers even further due to a change in feeding 2555 behavior (retention). Non-Prime males experience a release from competition after the pupation of the 2556 Prime male, and also benefit from the day 6 addition of food. The release from competition results in a 2557 0.01 mg increase in size (FxD), while the combined release and extra growth on the day 6 food results in 2558 a 0.14 mg increase (FxDxT). Unlike females, males appear to have two life history strategies: early 2559 pupation vs taking longer and growing larger. Early pupation may enhance mating success for males, 2560 while larger adults live longer and may have more opportunities to mate.

2561
Unlike females, the DxAxT and DxT interactions appear to produce size distributions of males that 2562 resemble the FxDxT and FxD competitive outcomes. All of the significant interactions that include 2563 density as one of the factors result in outcomes that look like competition among males, even those that 2564 do not include total food. The factors that describe the distribution of food in time (aliquot, timespan) 2565 interact with density independently of the food level to alter the competitive environment for males, 2566 but not for females. Note that the 4-way interaction (FxDxAxT) is not significant for males in the