Inhibition of testicular growth and development in Manduca sexta larvae parasitized by the braconid wasp Cotesia congregata

Tobacco hornworm larvae parasitized by the gregarious larval endoparasitoid Cotesia congregata exhibited an inhibition in testicular growth and development, the extent of which was determined by the age and developmental stage of the host at the time of parasitization. The degree of parasitic castration, as assessed by measurements of testicular volume, was correlated with the stadium in which parasitization occurred. A mathematical formula requiring the measurement of testicular length, width and depth was used to calculate testicular volume. The use of the depth parameter revealed a negative correlation between host weight and testicular volume in parasitized larvae. Testicular volumes of fifth instar hosts, which had been parasitized in the first stadium, were significantly smaller than those originally parasitized as fourth or fifth instar larvae and were not correlated with parasitoid load. Effects of natural parasitism were not duplicated by injections of C. congregata polydnavirus and venom, topical treatment with the juvenile hormone analog methoprene, or starvation of nonparasitized larvae. Larvae receiving virus plus venom or methoprene grew larger due to delayed wandering and had larger testes than controls. Deleterious effects on host testes may be due to the effects of nutrient competition between the developing parasitoid progeny and the gonads, combined with the juvenilizing effects believed to be caused by the polydnavirus.     

Effects of Glyptapanteles liparidis (Hym.: Braconidae) parasitism, polydnavirus, and venom on development of microsporidia-infected and uninfected Lymantria dispar (Lep.: Lymantriidae) larvae

Effects of parasitism, polydnavirus, and venom of the endoparasitoid Glyptapanteles liparidis on Lymantria dispar larvae infected with the microsporidium Vairimorpha sp. and uninfected hosts were studied. We tested the impact on growth and development of hosts, as well as on microsporidian infection. Both parasitism and polydnavirus/venom treatment alone caused a slight increase in growth rate and relative growth rate in uninfected fourth instar hosts. This effect was more pronounced with the addition of Vairimorpha infection. With no parasitism, however, infection reduced host growth markedly. Microsporidiosis delayed larval molts of L. dispar, and additional polydnavirus/venom treatment or parasitization induced significantly earlier molting. Polydnavirus/venom treatment of uninfected L. dispar resulted in prolonged larval development due to supernumerary molts and in higher pupal mortality. Infected larvae treated with polydnavirus/venom died earlier than infected larvae that were not treated and produced more Vairimorpha spores per unit fresh mass of the host.     

Parasitic castration of Plutella xylostella larvae induced by polydnaviruses and venom of Cotesia vestalis and Diadegma semiclausum

In the present study, we used gamma-ray to irradiate the female parasitoids to make wasp eggs infertile, resulting in pseudoparasitization, which allowed the analysis of maternal secretions such as polydnaviruses (PDVs) and venom in the absence of larval secretions or teratocytes by the growing parasitoids. We then investigated the spermatogenesis and components of testicular proteins of male Plutella xylostella larvae pseudoparasitized by two endoparasitoids (Cotesia vestalis and Diadegma semiclausum). The results showed that pseudoparasitism by the two endoparasitoids at the early third instar host larvae both induced smaller testes in size than those of nonparasitized host larvae. Both of them caused parasitic castration, and the degree of castration is almost as severe as in naturally parasitized hosts. This suggested that PDVs and venom played a major role in the degeneration of host testes. There are significant differences in the degree of castration induced by the two endoparasitoids, with respect to testicular growth, testicular protein concentrations, and histological changes of germ cells. Cotesia vestalis bracovirus always has a significantly stronger effect on host testicular growth and development than D. semiclausum ichnovirus. SDS-PAGE analysis indicated that synthesis of P 65 and P 67 proteins were clearly inhibited in testes of hosts that were pseudoparasitized by C. vestalis while reduction in synthesis of other proteins was not evident.     

Effect of parasitism on a nucleopolyhedrovirus amplified in Spodoptera frugiperda larvae parasitized by Campoletis sonorensis

We evaluated the consequences of parasitism by the solitary ichneumonid endoparasitoid Campoletis sonorensis(Cameron) towards the replication, genetic composition and virulence of a nucleopolyhedrovirus (Baculoviridae) originating from Spodoptera frugiperda(J. E. Smith) larvae. Parasitism by C. sonorensisand viral infection of third and fourth instar S. frugiperdalarvae resulted in reduced growth compared with nonparasitized control larvae. A positive correlation was observed between virus yield and larval instar at the moment of infection. When larvae were virus-inoculated in the fourth instar, parasitism resulted in a significant reduction in mean per capita virus yield compared to the virus yield from nonparasitized larvae. In an experiment involving 10 serial passages of virus in both parasitized and nonparasitized larvae, restriction endonuclease analysis of viral DNA amplified in nonparasitized larvae revealed the presence of the wild-type virus as well as three additional variants (A, B, and C) diagnosed by the presence of novel submolar PstI fragments of different sizes. In contrast, analysis of viral DNA from parasitized larvae showed the presence of the wild-type virus and two other variants (E and F), each characterized by a different submolar BglII fragment. Southern blot analysis indicated that the submolar fragments of variants E and F contained sequences originating from the viral genome. Bioassay of the different virus variants in S. frugiperdalarvae indicated that their virulence was equal or less than that of the wild-type virus. We conclude that parasitism can affect the quantity of virus produced in dually infected and parasitized larvae, but no adverse effects were detected in terms of the biological activity of the virus.     

Physiological and endocrine changes associated with polydnavirus/venom in the parasitoid-host system Chelonus inanitus-Spodoptera littoralis

As shown earlier, parasitization by the egg-larval parasitoid C. inanitus causes in its host the precocious onset of metamorphosis in the 5th instar followed by developmental arrest in the prepupal stage. Polydnavirus/venom were shown to be responsible for the developmental arrest. We investigated how polydnavirus/venom affect growth of the host larvae and found that head capsule widths were smaller from the 4th to 6th stadium and weights were lower in the 6th stadium in polydnavirus/venom-containing larvae than in non-parasitized larvae. In an attempt to identify endocrine parameters that are modified by polydnavirus/venom and might be responsible for the developmental arrest in the prepupa, we compared juvenile hormones, juvenile hormone esterase and ecdysteroids between non-parasitized and polydnavirus/venom-containing larvae from the 4th instar until pupation or developmental arrest, respectively. Obvious differences became manifest only in the 6th instar at the pupal cell formation stage, i.e. 12 days after entry of polydnavirus/venom into the host egg. Then, prothoracic glands of polydnavirus/venom-containing larvae released less ecdysteroids and ecdysteroid titres were lower than in non-parasitized larvae; this was followed by a delayed, reduced and desynchronized increase in prepupal juvenile hormones and juvenile hormone esterase and a slightly modified metabolism of ecdysone. This indicates that polydnavirus/venom affects the endocrine system of the host only after pupal commitment and that inhibition of prothoracic gland activity is the first detectable effect.     

Seasonal phenology and stage-specific parasitism of the apple ermine moth, Yponomeuta malinellus Zeller, in Korea

The apple ermine moth, Yponomeuta malinellus Zeller (Lepidoptera: Yponomeutidae), is a tent caterpillar that feeds on Malus spp. in Korea. Populations of the moth in native areas appeared to be regulated by the assemblage of parasitoids. Phenological associations between host stages and parasitoids, susceptible stage(s) of the host for each parasitoid, and stage‐specific parasitism were studied. The egg larval parasitoid Ageniaspis fuscicollis (Dalman) had highest parasitism of first instar larvae (24%), with 14% parasitism of other larval stages. Dolichogenidea delecta (Haliday) was recovered from all larval instars with the highest parasitism rate of second instar larvae (20.1%), followed by 19.9% parasitism of mid‐larval hosts. Herpestomus brunicornis Gravenhorst was reared from second instar larvae through to pupal collection, and had the highest parasitism rate (29.9%) at the pupal stage. The larval pupal parasitoid Zenillia dolosa (Meigen) was recovered from mid‐larval to pupal stages with the highest parasitism rate (5.5%) occurring in third to fourth instar larvae. The host stages for developing A. fuscicollis completely overlap with those of D. delecta, and with those of H. brunicornis to some degree. A statistically significant negative correlation exists between A. fuscicollis and these dominant parasitoids, indicating competitive interaction within the host.     

Parasitoid–Pathogen–Pest Interactions of Chelonus insularis, Campoletis sonorensis, and a Nucleopolyhedrovirus in Spodoptera frugiperda Larvae

In this study we examined interactions between two solitary endoparasitoids, the braconid Chelonus insularis and the ichneumonid Campoletis sonorensis, and a multiple-enveloped nucleopolyhedrovirus infecting Spodoptera frugiperda larvae. We examined whether ovipositing females minimize interference by discriminating amongst hosts and examined the outcome of within-host competition between parasitoid species and between the parasitoids and the virus. The egg–larval parasitoid Ch. insularis did not discriminate between virus-contaminated and uncontaminated S. frugiperda eggs; all S. frugiperda larvae that emerged from surface-contaminated eggs died of viral infection prior to parasitoid emergence. The larval parasitoid C. sonorensis also failed to discriminate between healthy and virus-infected S. frugiperda larvae or between larvae unparasitized or parasitized by Ch. insularis. Host larvae parasitized in the egg stage by Ch. insularis were suitable for the development of C. sonorensis when they were multiparasitized by C. sonorensis as first, second, third, and fourth instars, whereas emergence of Ch. insularis was dramatically reduced (by 85 to 100%) in multiparasitized hosts. Nonspecific host mortality was significantly higher in multiparasitized hosts than in singly parasitized hosts. The development time and sex ratio of C. sonorensis in multiparasitized host larvae were unaffected by the presence of Ch. insularis larval stages. Both Ch. insularis parasitized and nonparasitized larvae of the same instar (second, third, or fourth instars) had a similar quantitative response to a challenge of virus inoculum. All host larvae that ingested a lethal dose of virus were unsuitable for Ch. insularis development. In contrast, C. sonorensis did not survive in hosts that ingested a lethal virus dose immediately after parasitism, but parasitoid survival was possible with a 2-day delay between parasitism and viral infection and the percentage of parasitoid emergence increased significantly as the interval between parasitism and viral infection increased. The development time of C. sonorensis was significantly reduced in virus-infected hosts compared to conspecifics that developed in healthy hosts. C. sonorensis females that oviposited in virus-infected hosts did not transmit the virus to healthy hosts that were parasitized subsequently. Field applications of virus for biocontrol of S. frugiperda may lead to substantial mortality of immature parasitoids, although field experiments have not yet demonstrated such an effect.     

Overview of parasitism associated effects on host haemocytes in larval parasitoids and comparison with effects of the egg-larval parasitoid Chelonus inanitus on its host Spodoptera littoralis

In the first part we review the effects of larval endoparasitoids and their polydnavirus and venom on the immune system of their hosts. In all systems investigated, haemocyte spreading and encapsulation activity was reduced; in some cases effects on total (THC) or differential (DHC) haemocyte count as well as modification of haemocyte morphology and ultrastructure were also documented. In many cases polydnavirus (and venom) were shown to play a major role in abrogation of the host's immune reaction. In the second part we present the first investigation of effects of parasitism and polydnavirus/venom on the immune system of the host for an egg-larval parasitoid, Chelonus inanitus. We observed that in 4th and 5th instar larvae, i.e. 7 to 10 days after parasitization, neither haemocyte spreading and encapsulation activity, nor DHC, nor haemocyte ultrastructure were altered. After parasitization with X-ray irradiated wasps, which inject polydnavirus and venom and infertile eggs, there was no alteration of the above mentioned parameters. Nevertheless, parasitoid larvae implanted into 4th instar larvae which developed from eggs parasitized with X-ray irradiated wasps were not encapsulated, whereas co-injected latex beads were. These results show that parasitism by this egg-larval parasitoid does not generally suppress the host's immune system but that polydnavirus/venom injected at oviposition prevent, by, as yet unknown mechanisms, encapsulation of the parasitoid larva.     

Possible bases of pseudoparasitism in Spodoptera littoralis larvae stung by Microplitis rufiventris

The effects of host age and parasitoid female age on the occurrence of 'Pseudoparasitism', using the Spodoptera littoralis-Microplitis rufiventris host-parasitoid system were investigated. The first four larval instars of the host are not equally suitable for parasitoid development. The proportion of pseudoparasitized hosts significantly increases when: (1) the age of the female parasitoid increases; (2) oviposition occurs mostly in fourth instar larvae; (3) a later age of the host instar is used; (4) the mandibles of the newly hatched parasitoid larvae mistakenly attack host interior organs (e.g. Malpighian tubules); and (5) an imperfect growth pattern of teratocytes occurs. The reluctance of female wasps to parasitize fourth instar host larvae is not due to the thickness of host cuticle but possibly due to the unfavourable physiological state of the host larvae. The age of host larvae at the time of parasitization may influence the adverse effects of parasitoid factors (e.g. polydnavirus, venom and teratocytes) on the growth of host larvae. It is suggested that females of M. rufiventris are able to determine the suitability of a potential host instar for the development of their offspring. The cell diameter of M. rufiventris teratocytes increases with increasing age of host larvae at the time of oviposition. The association within the host of living parasitoid larvae and functional teratocytes may be important for the survival of each other and consequently for successful parasitism.     

Stimulated synthesis of the female-specific storage protein in male larvae of the silkworm Bombyx mori treated with juvenile hormone analog

Application of methoprene to fourth (penultimate) instar larvae of the silkworm Bombyx mori induced the appearance of the feeding dauer larvae at the fifth (last) instar and prevented pupal metamorphosis. Methoprene also increased the protein concentrations of hemolymph last instar larvae by preventing sequestration of storage proteins by the fat body. Usually, the female-specific storage protein 1 (SP1)* disappears from the male hemolymph at the time of the last larval instar. However, exposure of male larvae to methoprene at the penultimate instar enhanced the accumulation of SP1 in the hemolymph. The SP1 accumulated in males did not differ in molecular weight and immunoreactivity from the SP1 produced in female larvae. Both sexes of fourth instar larvae allatectomized on day 1 instantly accumulated SP1 in the hemolymph, and methoprene application after allatectomy suppressed the hemolymph accumulation of the SP1. In contrast, if allatectomy was carried out at a later stage of the fourth larval instar, SP1 concentration in hemolymph of fifth instar larvae did not increase, suggesting the different juvenile hormone action for regulation of SP1 synthesis in the penultimate instar larvae of silkworms.     

Parasitic castration of Pseudaletia separata by Cotesia kariyai and its association with polydnavirus gene expression

Parasitization by the endoparasitoid Cotesia kariyai caused the inhibition of spermatogenesis of Pseudaletia separata. This phenomenon is called parasitic castration. The degree of castration was dependent on the host stage parasitized. Host parasitized on day 1 of the 4th stadium (the time of primary spermatocyte accumulation), had testicular cells with abnormal chromosomes appearing two days after parasitization, and spermiogenesis was completely inhibited. However, when hosts were parasitized on day 0 of the 6th (final) stadium, the degree of castration was less severe, and elongated cells appeared similar to those found in nonparasitized larvae. Results of this study involving injection of C. kariyai polydnavirus (CkPV) and venom suggested that these wasp components caused the appearance of abnormal chromosomes in specific germ cells, which were in mitotic or meiotic prophases. The amount of CkPV gene expression in host testes increased immediately after parasitization and reached a maximum 12h later. The early-expressed CkPV gene(s) may be related to the parasitic castration phenomenon.     

The endoparasitoid Cotesia kariyai (Ck) regulates the growth and metabolic efficiency of Pseudaletia separata larvae by venom and Ck polydnavirus

It was previously demonstrated that parasitization by Cotesia kariyai caused a decrease in weight gain and food consumption in host larvae, resulting in a lower final weight for parasitized hosts. It is predicted that C. kariyai regulates the physiological condition of the host to obtain maximum food under restricted nutritional conditions. Approximate digestibility (AD) was higher following parasitization but the efficiency of conversion of digested food (ECD) of the parasitized hosts was lower. This suggests that resources available to the parasitoid larvae are enhanced in the parasitized hosts. We evaluated the physiological changes caused by injection of calyx fluid (polydnavirus) plus venom (C+V) in nonparasitized hosts. Injection of C+V into the nonparasitized hosts duplicated the effects of parasitism, namely it increased the AD and decreased the ECD. Furthermore, C+V injections elevated trehalose concentrations in nonparasitized host 7 to 10 d after injection (2nd stadium of the parasitoid larva). Protein content also increased on days 9 and 10 after C+V injection. These results suggest that the nutrients that parasitoid larvae require for their growth increase in the hemolymph of the host during the 2nd stadium of the parasitoid larva.     

The effect of host developmental stage at parasitism on sex‐related size differentiation in a larval endoparasitoid

• 1 For their larval development, parasitoids depend on the quality and quantity of resources provided by a single host. Therefore, a close relationship is predicted between the size of the host at parasitism and the size of the emerging adult wasp. This relationship is less clear for koinobiont than for idiobiont parasitoids.
• 2 As size differentiation in host species exhibiting sexual size dimorphism (SSD) is likely to occur already during larval development, in koinobiont larval endoparasitoids the size of the emerging adult may also be constrained based on the sex of the host caterpillar.
• 3 Sex‐specific growth trajectories were compared in unparasitised Plutella xylostella caterpillars and in second and fourth instar hosts that were parasitised by the solitary larval koinobiont endoparasitoid Diadegma semiclausum. Both species exhibit SSD, where females are significantly larger than males.
• 4 Healthy female P. xylostella caterpillars developed significantly faster than their male conspecifics. Host regulation induced by D. semiclausum parasitism depended on the instar attacked. Parasitism in second‐instar caterpillars reduced growth compared to healthy unparasitised caterpillars, whereas parasitism in fourth‐instar caterpillars arrested development. The reduction in growth was most pronounced in hosts producing male D. semiclausum.
• 5 Parasitism itself had the largest impact on host growth. SSD in the parasitoid is mainly the result of differences in growth rate of the parasitoid–host complex producing male and female wasps and differences in exploitation of the host resources. Female wasps converted host biomass more efficiently into adult biomass than males.

     

Inhibition of juvenile hormone esterase activity in Lymantria dispar (Lepidoptera, Lymantriidae) larvae parasitized by Glyptapanteles liparidis (Hymenoptera, Braconidae)

Glyptapanteles liparidis is a gregarious, polydnavirus (PDV)-carrying braconid wasp that parasitizes larval stages of Lymantria dispar. In previous studies we showed that parasitized hosts dramatically increase juvenile hormone (JH) titers, whereas JH degradation is significantly inhibited in the hemolymph. Here we (i) quantified the effects of parasitism on JH esterase (JHE) activity in hemolymph and fat body of penultimate and final instars of L. dispar hosts and (ii) assessed the relative contribution of individual and combined wasp factors (PDV/venom, teratocytes, and wasp larvae) to the inhibition of host JHE activity. The effects of PDV/venom was investigated through the use of gamma-irradiated wasps, which lay non-viable eggs (leading to pseudoparasitization), while the effects of teratocytes and wasp larvae were examined by injection or insertion of these two components in either control or pseudoparasitized L. dispar larvae. Parasitism strongly suppressed host JHE activity in both hemolymph and fat body irrespective of whether the host was parasitized early (premolt-third instar) or late (mid-fourth instar). Down-regulation of JHE activity is primarily due to the injection of PDV/venom at the time of oviposition, with only very small additive effects of teratocytes and wasp larvae under certain experimental conditions. We compare the results with those reported earlier for L. dispar larvae parasitized by G. liparidis and discuss the possible role of JH alterations in host development disruption.     

Biology of Microplitis kewleyi (Hymenoptera: Braconidae): A parasite of Agrotis ipsilon (Lepidoptera: Noctuidae) larvae

Microplitis kewleyi Muesebeck is a gregarious internal parasite of larvae of the black cutworm Agrotis ipsilon (Hufnagel). Studies of the biology of the parasite revealed that there was an inverse relationship between host instar and parasite preference. Duration of development from egg to pupa ranged from 18 days at 27°C to 68.7 days at 16°C. Development from egg to pupa took 13.5–21.6 days when fourth and first instar host larvae, respectively, were parasitized. A larger number of parasites emerged from hosts parasitized in the fourth instar (22.4) than the first instar (11.5). Parasite pupation occurred when the host was in the fifth/sixth instar, depending on the instar parasitized. Thirty‐nine per cent of host larvae exposed as first instars to parasites died before parasite emergence. This decreased to 0% for host larvae exposed as fourth instars. The sex ratio was 1:1.2 (M:F). Thirty‐seven per cent of hosts exposed diurnally were stung, compared to 24% exposed nocturnally. Mean daily progeny was highest (12) on the first day, decreasing to zero after 20 days. Percent host parasitism was also highest on the first day (35%) decreasing to nearly 0% after 18 days. There appear to be three parasite larval instars. Host larvae often remained alive after parasite emergence.     

Teratocytes of the solitary endoparasitoid Meteorus gyrator (Hymenoptera: Braconidae): morphology, numbers and possible functions

Abstract. Laboratory studies investigated the development of teratocytes derived from the eggs of the parasitoid Meteous gyrator (Thun.) in its host, the tomato moth Lacanobia oleracea (L.). At hatching, each parasitoid egg produced an average of approximately 1000 teratocytes, but this number declined to approximately 400 during the course of parasitism. The teratocytes increased in size markedly, such that 7 days after egg hatch their mean diameter was approximately four times that of the cells immediately after dissociation. The haemolymph of parasitized hosts had reduced phenoloxidase activity, and teratocytes inhibited phenoloxidase activity when coincubated with plasma from nonparasitized hosts. The injection of teratocytes into nonparasitized fifth‐instar L. oleracea larvae suppressed growth and induced a supernumerary moult in some larvae. A number of parasitism‐specific proteins were detected in the haemolymph of parasitized hosts, and incubation of teratocytes in culture media indicated that these cells were a source of at least two of these proteins.     

Stage dependent influences of polydnaviruses and the parasitoid larva on host ecdysteroids

In the solitary egg-larval parasitoid Chelonus inanitus (Braconidae) both polydnavirus and the parasitoid larva manipulate host development. Parasitization leads to a premature drop in juvenile hormone titre and a precocious onset of metamorphosis in the 5th larval instar. The C. inanitus bracovirus (CiBV) alone causes a reduction in host ecdysteroid titres at the pupal cell formation stage and prevents pupation. Here we report three new findings. (1) We show that parasitization causes a reduction in haemolymph ecdysteroid titre immediately after the moult to the 5th instar; similarly low values were seen in nonparasitized larvae after the moult to the 6th instar. These data along with parasitoid removal experiments indicate that the low ecdysteroid titre after the moult is a very early sign of the upcoming metamorphosis. (2) In vitro experiments with prothoracic glands and brain extracts showed that CiBV affects both prothoracic glands and prothoracicotropic hormone after the stage of pupal cell formation. (3) In the haemolymph of parasitized larvae the ecdysteroid titre increased in the late cell formation stage, i.e. immediately before egression of the parasitoid. In vitro experiments showed that late 2nd instar parasitoids release ecdysteroids and are thus very likely responsible for the rise in host ecdysteroids.     

Transient expression of specific Cotesia plutellae bracoviral segments induces prolonged larval development of the diamondback moth, Plutella xylostella

A polydnavirus, Cotesia plutellae bracovirus (CpBV), possesses a segmented and dispersed genome that is located on chromosome(s) of its symbiotic endoparasitic wasp, C. plutellae. When the host wasp parasitizes larvae of the diamondback moth, Plutella xylostella, at least 27 viral genome segments are delivered to the parasitized host along with the wasp egg. The parasitized P. xylostella exhibits significant immunosuppression and a prolonged larval development. Parasitized larvae take about 2 days longer than nonparasitized larvae to develop until the wandering stage of the final larval instar, and die after egress of the full grown wasp larvae. Developmental analysis using juvenile hormone and ecdysteroid analogs suggests that altering endocrine signals could induce the retardation of larval developmental rate in P. xylostella. In this study we used a transient expression technique to micro-inject individual CpBV genome segments, and tested their ability to induce delayed larval development of P. xylostella. We demonstrated that a CpBV segment was able to express its own encoded genes when it was injected into nonparasitized larvae, in which the expression patterns of the segment genes were similar to those in the larvae parasitized by C. plutellae. Twenty three CpBV genome segments were individually cloned and injected into the second instar larvae of P. xylostella and their effects assessed by measuring the time taken for host development to the cocooning stage. Three CpBV genome segments markedly interfered with the host larval development. When the putative genes of these segments were analyzed, it was found that they did not share any common genes. Among these segments able to delay host development, segment S27 was predicted to encode seven protein tyrosine phosphatases (CpBV-PTPs), some of which were mutated by insertional inactivation with transposons, while other encoded gene expressions were unaffected. The mutant segments were unable to induce prolonged larval development of P. xylostella. These results suggest that CpBV can induce prolonged larval development of P. xylostella, and that at least some CpBV-PTPs may contribute to the parasitic role probably by altering titers of developmental hormones.