Descriptions of immature stages of four species of the genera Graptus, Peritelus, Philopedon, and Tanymecus and larval instar determination in Tanymecus (Coleoptera, Curculionidae, Entiminae)

Abstract The mature larva and pupa of Graptustriguttatustriguttatus and the mature larva of Peritelussphaeroides are described for the first time. The larvae of Philopedonplagiatum and Tanymecuspalliatus are re-described. Five larval instars were determined in Tanymecus, thereby correcting doubtful data in the literature. The relationship between larval growth, number of larval instars, head width of the mature larva, and the adult weevil is explained using the example of Tanymecus. The nearly constant ratio of subsequent larval instars in head width ratio, termed “growth factor” and derived from Dyar’s ratio, is used for the determination of larval instars. Larval collecting and breeding data are discussed in relation to their significance for the clarification of life-cycles.


Introduction
In this continued contribution on larvae of the subfamily Entiminae Schönherr, 1823 we describe or redescribe the mature larvae of four further species (e.g., Sprick and Gosik 2014;Gosik et al. 2016Gosik et al. , 2017 and the pupa of one species. They represent four different tribes: Byrsopagini Lacordaire, 1863 (= Alophini LeConte, 1874), Cneorhinini Lacordaire, 1863, Peritelini Lacordaire, 1863, and Tanymecini Lacordaire, 1863(Alonso-Zarazaga et al. 2017. They also allow some insight into the morphological diversity of Central European Entiminae larvae. In the present paper we describe for the first time the mature larva and pupa of Graptus triguttatus (Fabricius, 1775). For this species, Van Emden (1952) provided a description of the first instar larva, eggs and oviposition habit. Dudich (1921) provided two host plant records, Beta vulgaris L. and Symphytum officinale L., and some data about oviposition and egg morphology, but no relevant information about larval or pupal stages was given. For Philopedon plagiatum (Schaller, 1783) there is a description of the mature larva by Van Emden (1952), but only the head capsule and the right pedal lobe were illustrated. A study of distribution and biology of this species in Great Britain was published by Morris (1987).
From the tribe Peritelini we describe the mature larva of Peritelus sphaeroides Germar, 1824 for the first time. The pupa was already described by Gosik and Sprick (2013). We do not know of any description of a Central European species in this tribe. Van Emden (1952) and Rosenstiel (1987) characterized the larvae of two North American Peritelini genera, Nemocestes Van Dyke, 1936 andPeritelinus Casey, 1888. Despite the frequency and abundance of Tanymecus palliatus (Fabricius, 1787), and the good characterization of its development (Hoffmann 1963;Dieckmann 1983), there is no detailed description of the larval instars of this species. Only Znamenskij (1927), in his keys to soil insects, depicted the habitus and last abdominal segment. But this source is not readily available, and we received only a few pages of this work through the kindness of Vitaliy Nazarenko. These studies had been carried out after damage by this species to sunflower and beet fields in Ukraine and southern Russia in the 1920s and 1930s. To our knowledge, the most complete description of a Tanymecus Germar, 1817 larva was published by Catrinici (1944) for T. dilaticollis. It is supplemented by Van Emden (1952) with a description of the mature larva of T. palliatus (without figures) and a description of the North American T. confusus ("or very near"). The pupa of T. palliatus was already described by Gosik and Sprick (2013).
The aim of this paper is to describe the mature larvae of the four Entiminae species mentioned before and to give some examples about how to use data from larval descriptions for the determination of larval instars and for the study of life-cycles. An important prerequisite for studying life-cycles is to have correctly identified larvae, which is often difficult and a main reason why life-cycles of Entiminae weevils, apart from some noxious Otiorhynchus and Sitona species, are usually little known.

Materials
Specimens of three of the four species studied were collected in the field under certain plants and usually at the same sites where adults were previously collected. Larval instars of the fourth species, Peritelus sphaeroides Germar, 1824, were obtained in captivity by breeding in an air-conditioned room (see Gosik et al. 2016). Two searches for preimaginal stages at the field site where adults of this species were known to occur, were unsuccessful. Number of specimens examined, date and places of collecting are given ahead of the description of each species. As "mature" we regard the larvae with the largest head capsule widths (most closely corresponding to head size of pupa and adult of the species). We also take into consideration results of measurements (if available) provided by other authors.

Methods
All specimens studied were fixed in 75% ethanol and examined under an optical stereomicroscope (Olympus SZ 60 and SZ11) with calibrated oculars. Measurements of larval instars were made for: body length (BL), body width (BW) (usually at abdominal segment I or II), width (HW) and height (HH) of the head capsule (see Fig. 18). In pupae, body length (BL), body width (BW) (at the level of middle legs) and width of pronotum (= thorax) (THW) were measured.
The observations and measurements were conducted using a light compound microscope with calibrated oculars. Drawings and outlines were made using a drawing tube (MNR-1) installed on a stereomicroscope (Ampliwal) and processed by computer software (Corel Photo-Paint X7, Corel Draw X7). Photos were taken with an Olympus E-M10 or using an Olympus BX63 microscope and processed by Olympus cellSens Dimension software. The larvae selected for pictures using SEM (scanning electron microscope) were at first dried in absolute ethyl alcohol (99.8%), rinsed in acetone, treated by CPD procedure (critical point drying) and then gold-plated. For the examination of selected structures a TESCAN Vega 3 SEM was used. General terminology and chaetotaxy follow Anderson (1947), May (1994), Marvaldi (1997Marvaldi ( , 1998aMarvaldi ( , 1998bMarvaldi ( , 1999 and Skuhrovec et al. (2015), with terminology for antennae following Zaharuk (1985), May (1994) and Marvaldi (1998).
We follow Trnka et al. (2015) and Skuhrovec et al. (2015) who counted in weevils 3 pairs of ams and 2 pairs of mes. Position of the distal pair of mes is still questionable and some other authors (e.g. May 1994;Marvaldi 1998) reported for weevil larvae 2 pairs of ams and 3 pairs of mes, and they regarded ams 1 as mes 3 .
All these specimens are deposited in the collection of the Department of Zoology, Maria Curie-Skłodowska University (Lublin, Poland). In Table 4 the chaetotaxy of the larvae is given. If necessary, head width of adults was measured directly behind eyes.
Larval instar determination is based on Dyar's law (Dyar 1890), which had been developed and refined by Leibee et al. (1980), Rowe and Kok (1985) and Sprick and Gosik (2014). For the instar determination only data of L 1 larvae and of mature larvae are needed. We explain in several steps how this method works, define the 'growth factor' (based on Dyar's ratio), use it in all detail in Tanymecus dilaticollis and show how to find the best approximation of the factor that determines larval growth. Body (Figs 1-3). Moderately slender, curved, rounded in cross section. Prothorax slightly narrower than mesothorax; metathorax as wide as mesothorax. Abdominal segments 1-6 of almost equal length; 7-9 decreasing gradually to the terminal body part; 10 reduced to 4 anal lobes with the largest in dorsal and the smallest in ventral position, lateral lobes of equal size (Fig. 3). Spiracles (of thoracic and abdominal segments  1-8) annular with 2 vestigial airtubes. Chaetotaxy well developed, setae capilliform, variable in length, dark yellow to brown. Each side of prothorax ( Fig. 13) with 8 prns of unequal length: 5 of them placed on the weakly visible premental sclerite, next 3 short setae close to spiracle; 2 ps and 1 eus. Meso-and metathorax (Fig. 13) on each side with 1 short prs, 4 pds, variable in length: first, third and fourth long, second very short, 1 1-3 -thoracic segments 1-3, Abd. 1-10 -abdominal segments 1-10, setae: as -alar, ps -pleural, epsepipleural, ds -dorsal, lsts -laterosternal, eus -eusternal, pda -pedal, pds -postdorsal, prns -pronotal, prs -prodorsal, ss -spiracular, sts -sternal, ts -terminal. short as, 3 short ss, 1 moderately long eps, 1 moderately long ps and 1 eus. Each pedal area of thoracic segments with 6 pda, variable in length. Abd. 1-7 (Figs 14-17) on each side with 1 short prs, 5 pds variable in length(first, third and fifth long, second and fourth short) and arranged along the posterior margin of each segment, 1 minute and 1 short ss, 2 eps and 2 ps of various length, 1 lsts and 2 short eus.  on each side with 1 short prs, 4 pds variable in length (first and third moderately long, second and fourth long) and arranged along the posterior margin of the segment, 1 minute ss, 2 eps and 2 ps of various length, 1 lsts and 2 short eus. Abd. 9 (Figs 15-17) on each side with 3 ds (dorsal setae), first moderately long, second and third long, all located close to the posterior margin of the segment, 1 long and 1 minute ps and 2 short sts. Each lateral anal lobe (Abd. 10) with a pair of minute setae. Head (Fig. 18). Light to dark yellow, oval, frontal suture distinct, Y-shaped, endocarina present, reaching to middle of frons. Setae on head capilliform; des 1, 2, 3, 5 equal in length; des 1 and des 2 located in the central part of epicranium, des 3 placed on frontal suture, des 5 located anterolaterally; fs 4, 5 equal in length, fs 4 located anteromedially, fs 5 anterolaterally, close to epistome; les 1 and les 2 equal in length, less than half the length of des 1 ; ves short, poorly developed. Postepicranial area with 3 very short pes (Fig. 18). Two weakly visible stemmata close to des 5 . Antennae (Fig. 19) located at the end of frontal suture; antennal segment membranous, bearing sensorium (Se) conical, almost as wide as long, located medially, and 6 sensilla of different types: 1 sa and 5 sb. Labrum (Fig. 20) almost semicircular, anterior margin rounded; 3 pairs of lrs, different in length, lrs 1 and lrs 2 very long, lrs 3 moderately long; lrs 1 placed medially, lrs 2 anteromedially, lrs 3 anterolaterally. Clypeus ( Fig. 20) trapezoid, its anterior margin slightly concave, covered with asperities; 2 pairs of cls short, located posteromedially; clss clearly visible, placed medially between cls. Epipharynx ( Fig. 20) with 3 pairs of finger-shaped als of almost equal length; 3 pairs of ams: ams 1 and ams 3 rod-shaped, very short, ams 2 finger-like, very long; 2 pairs of rod-shaped mes of various lengths: first pair placed medially, second pair anteriorly, very close to ams. Surface of epipharynx smooth. Labral rods elongate, converging posteriorly. Mandibles ( Fig. 21) curved, narrow, with slightly divided apex (teeth of various lengths). There is an elongate protuberance on the cutting edge between the apex and the middle of the mandible; both mds capilliform, different in length, placed transversely. Maxilla  with 1 stps and 2 pfs of equal length; mala with 7 finger-or rod-like dms of almost equal size, 4 vms , varied in length and all shorter than dms; mbs short. Maxillary palpi with 2 palpomeres, basal with short mps; distal palpomere apically with a group of sensilla, each palpomere with a pore. Basal palpomere distinctly wider than distal, both of almost equal length. Prelabium (Fig. 24) cup-like with 1 moderately long prms, located medially. Ligula with 3 pairs of minute ligs. Premental sclerite clearly visible, trident-shaped, posterior extension with acute apex. Labial palpi 2-segmented; apex of distal palpomere with some sensilla; each palpomere with  a pore. Basal palpomere distinctly wider than distal, both of almost equal length. Postlabium ( Fig. 24) with 3 capilliform pms (postlabial setae), the first pair located anteromedially, the remaining 2 pairs posterolaterally; pms 1 and pms 3 very short, pms 2 twice as long as others.  Description of the pupa. Body length (♂, ♀): 7.5-9.0 mm; body width (at level of mesocoxae): 3.8-4.5 mm; width of thorax: 2.0-2.3 mm.
Chaetotaxy well developed, setae variable in lengthand shape: spine-like or capilliform, dark yellow to brown, usually located on visible protuberances. Head capsule and rostrum include 1 vs, 2 minute sos, 1 spine-like and 1 minute os, 2 pas, 3 rs of varied sizes and 1 minute es. Except sos and es, all setae of the head and rostrum are placed on protuberances. Pronotum with 2 as, 1 ls, 2 ds and 2 pls. All setae of pronotum spine-like, of equal size (only ds 1 slightly larger than others); all setae placed on protuberances. Mesothorax with 2 minute setae placed anteromedially and 3 spine-like setae placed medially. Metathorax with 4 spine-like setae placed medially. Abdominal segments 1-7 with 7 pairs of d 1-7 : d 1-6 short, spine-like, placed on protuberances, in lines along the posterior margin of segments, d 7 short, capilliform, placed anterolaterally, and 2 minute l 1-2 . Setae no. 3 and no. 5 increasing gradually from segment 2 to 7. Segment 8 with 4 pairs of spine-like setae of varied lengths (d 1-4 ), placed on protuberances, in lines along the posterior margin of the segment. Seta no. 2 distinctly larger than others. Segment 9 with 3 pairs of short, capilliform v 1-3 . Each apex of femora with 2 fes, spine-like and of various length.

Peritelus sphaeroides
Specimens examined. Rearing was started on 02.05.2012 in the climate chamber of JKI in flowerpots with mainly Euonymus fortunei (Turcz.) Hand.-Mazz. and one with Prunus laurocerasus L. Adults had been collected 5 days previously in a hedgerow with ornamental shrubs in the JKI area.
Body (Figs 7-9). Slender, elongate, slightly narrowed bilaterally dorso-ventrally. Prothorax slightly smaller than mesothorax; metathorax as wide as mesothorax. Abdominal segments 1-7 of almost equal length. Abdominal segment 8 wide, flattened posteriorly, with conical lateral lobes. Abdominal segment 9 strongly reduced, consisting of 4 well-isolated lobes, ventral almost rounded, lateral conical, dorsal semicircular. Abdominal segment 10 consists of 4 anal lobes of almost equal size. Anus located ter-  minally, covered by lobes of abdominal segment 9. Apical parts of lateral lobes of the segments 6-8 and all lobes of segment 9 darkly sclerotized (Figs 46-48). Spiracles (of thoracic and abdominal segments 1-8) annular. Chaetotaxy well developed, setae capilliform, variable in length, yellowish to brown. Each side of prothorax ( Fig. 44) with 9 prns of unequal length, placed on the weakly sclerotized pronotal sclerite; 2 ps and 1 eus very short. Meso-and metathorax (Fig. 44) on each side with 1 moderately long prs and 4 pds, variable in length (first, second and fourth short, third moderately long), 2 short as, 3 minute (various in length) ss, 1 moderately long eps, 1 short ps and 1 eus. Each pedal area of thoracic segments with 9 pda, variable in length. Abd. segment 1-8 (Figs 45-48) on each side with 1 short prs and 5 pds, almost equal in length, arranged along the posterior margin of each segment, 1 minute and 1 long ss (segment 8 with 1 minute ss only), 4 eps (segment 6 with 3 eps, segments 7 and 8 with 2 eps) and 2 ps, equal in length, 1 lsts and 2 short eus. Abdominal setae increase slightly and gradually from segment 1 to 8. Abd. segment 9 (Figs 46-48) on each side with 2 moderately long ds, located near the posterior margin of the segment, 1 moderately long ps and 2 short sts. Anal lobes without setae.

The number of larval instars in Tanymecus
There are some strange statements about the number of larval instars in the larval stage of species of genus Tanymecus. Hoffmann (1963), who relied on authors from the former Soviet Union, reported about 10 larval instars in T. palliatus, which was already commented by Dieckmann (1983) as a 'for weevils surprising fact'. In T. dilaticollis Gyllenhal, 1834, Catrinici (1944 determined six larval instars. She reported that larval head width increased up to the fourth larval instar, decreased in the fifth and increased again in the sixth instar to nearly the same value as in the fourth. This sounds really strange and has to be taken with caution and tested with new observations. This was also the reason for Van Emden (1952) to propose four larval instars for T. dilaticollis.
For the exact determination of the number of larval instars we summarized and assessed our own measuring data and added data from literature, if necessary (Tables 1, 2).
Due to the dubiousness of the number of larval instars in T. dilaticollis given by Catrinici (1944) and Van Emden (1952) we used measuring data for the head width (HW) of adults of both species and of mature larvae of T. palliatus to assess the HW of the mature larva of T. dilaticollis. This ratio should be rather similar in two species of the same genus. Hence, the value calculated in this way for the HW of the mature larva of T. dilaticollis is 1.51 mm.
We also needed to determine the number of larval instars for both species: there are data for L 1 and for mature larvae, and in T. dilaticollis there are also measurements for several instars, even if (especially in the higher instars) the data are doubtful.
The determination of larval instars is mainly based on the method of Dyar (1890) and has been used by several authors, even if apparently not known to all scientists who have dealt with larvae. There are several publications about weevil larvae where this Table 1. Head width measuring data of the species studied. Results in mm; n -number of specimens measured, in adults behind eyes; L 1 -first instar larva; ML -mature larva; *: an assignment to this instar is doubtful. A transfer to 'mature larvae' would change the average value only slightly; **: data from Gosik and Sprick (2013). Data from literature in italics.

Species
Larval instars Pupa Adult L 1 larvae Premature larvae Mature larvae Graptus triguttatus (L 1 data from Van Emden 1952)   ratio was applied. We preferred to use Dyar's ratio -1 and called it Growth Factor (GF) as it corresponds more to the natural development.
In Mitoplinthus caliginosus (Fabricius, 1775) (subfamily Molytinae), after comparison of the growth factors 1.35, 1.4 and 1.5, the best approximation was found with a value of 1.4 for head capsule width (Sprick and Gosik 2014). This value agrees with Dyar's ratio of 0.714. Rowe and Kok (1985) gave a Dyar's ratio value for Rhinocyllus conicus (Frölich, 1792) (subfamily Lixinae) larvae of 0.65 (this agrees with a GF of 1.538). In agreement Leibee et al. (1980) determined the ratios of each instar of two populations of Sitona hispidulus (Fabricius, 1777) (subfamily Entiminae, tribe Sitonini) and reported Dyar's values between 0.642 and 0.739. The median value of these data is by our calculation 0.6995 (GF = 1.43). These data show that there are rather different values for larval growth and that there are also differences between the growth of different larval instars.
For larval instar determination in Tanymecus we tested four values between 1.4 and 1.5 to achieve the best approximation of larval growth. We started with the L 1 larva that we received from egg-laying of adult weevils (head width 0.38 mm) and calculated the subsequent instars with the selected GF values until 1.71 mm, the head width of the mature larvae, was achieved. For this procedure, five steps were needed. Higher GF values, as for example 1.538 in Rhinocyllus conicus, were excluded because of the reduced number of larval instars in this rather distantly related subfamily (Table 3).
From Table 2 it can easily be seen that both species have 5 larval instars. The best approximation is achieved with a GF of 1.44 in Tanymecus dilaticollis and 1.46 in T. palliatus (i.e. Dyar's ratio of 0.694 and 0.685, respectively). The small difference may be due to the absent HW variation of the two available adult T. dilaticollis specimens that showed both the same value and hence do not represent the HW variation of the population. Furthermore it can be stated that the values of Catrinici (1944) are beginning to seem doubtful from the fourth larval instar onward.
For this approximation it is only necessary to know the head width of the L 1 larva and that of the last instar. And the HW of the last instar can be assessed from the HW of the adult weevil, as it is shown in Table 2. In adults HW was always measured directly behind the eyes to avoid an excessive importance of prominent eyes, which could be a problem in genera such as Strophosoma Billberg, 1817 (see Gosik et al. 2017) or in species such as Tanymecus dilaticollis. Larval growth, number of larval instars and size of the adults' head width (and therefore size of adults, too) are in a very close relationship to each other. The same may be true for the HW of the pupa. An instar determination is also possible for Graptus triguttatus. According to Van Emden (1952), the head capsule width of the L 1 larva is 0.34 mm (average of three larvae; Table 1). The application of a GF of 1.45 shows a good approximation with the measuring values given in Table 1: 0.34 mm × 1.45 (repeatedly) = 0.493 mm (L 2 ), 0.715 mm (L 3 ), 1.037 mm (L 4 ) and finally 1.503 mm (L 5 ). Thus, Graptus triguttatus has also 5 larval instars, and the premature larvae from Table 1 may represent L 2 and L 4 larvae. There is a great variation in adults' head width in this species ranging from 1.05 mm to 1.5 mm (Table 1). This agrees with the span between the last larval instars and nearly achieves the supposed growth factor of 1.45, so that the assignment of a certain larva to the right instar is doubtful in extremely sized specimens, the more as head width variation becomes larger with instar and size. A similar size variation was observed in Peritelus sphaeroides adults and Philopedon plagiatum larvae (Table 1). Table 3. Larval instar determination for Tanymecus dilaticollis and T. palliatus. All measuring data in mm; initial data bold, calculated data in italics; target data of the approx imation bold and in italics.

Tanymecus palliatus
Growth factor (to be tested) In Peritelus sphaeroides and Philopedon plagiatum an instar determination is impossible due to the absence of L 1 head width data. It can only be concluded from the data for premature larvae in Philopedon plagiatum ( Table 1) that these data represent the penultimate instar. Opposite to Graptus and Tanymecus, the HW of Philopedon adults is greater than in mature larvae. In Peritelus sphaeroides the HW of adults is slightly smaller than in mature larvae, but not significantly.

I. Significance of morphological features of larvae for the relationship between genera and higher taxa
Philopedon belongs to genera with abdominal type 'B' larvae together with Strophosoma and Tanymecus ( Van Emden 1950, 1952. The main feature is the presence of a flattened dorso-ventral abdominal segment 9. Furthermore this feature (unique among weevils) may suggest an unknown kind of relationship between these genera (Gosik et al. 2017). Nevertheless, there are some morphological differences between larvae of type B, for example, the abdominal segment 10 is almost covered by segment 9 in Philopedon, whereas segment 10 is integrated with segment 9 and forms a sclerotized ventral surface in Strophosoma and Tanymecus. The dorsal lobe of segment 9 is largest in Strophosoma, whereas the ventral lobe is largest in Tanymecus.
According to Van Emden (1952), the chaetotaxy of the mature larva of Tanymecus palliatus is only slightly different from L 1 , namely: only six setae instead seven on the pedal lobe and lack of microseta des 4 . It is worth mentioning that the proportion between (and arrangement of ) setae shows the same order on L 1 as on mature larvae, a finding that is not common in Entiminae (Fidler 1936;Gosik and Sprick 2012b;Gosik et al. 2016Gosik et al. , 2017. To the characters listed by Van Emden (1952) as typical for the genus Tanymecus we can add that the clypeus is twice as long as labrum (Fig. 66).
The mature larva of Philopedon plagiatum is described by Van Emden (1952). The only exception is cls: rather short on L 1 versus extremely long on the mature larva; cls as long as lrs 1 is not observed in other Entiminae.
In Graptus triguttatus the shape of the body, number and proportion of setae on L 1 (according to Van Emden 1952) and of the mature larva (present paper) are almost identical. The tribe Byrsopagini (with the genus Graptus Schönherr, 1823) is currently removed from the subfamily Molytinae Schönherr, 1823 to the subfamily Entiminae Schönherr, 1823 (Thompson 1992;Marvaldi 1997Marvaldi , 1998aAlonso-Zarazaga and Lyal 1999;Löbl and Smetana 2013). From two apomorphies, most typical for Entiminae (4 vms and cushion-like sensorium) in the sense of Marvaldi (1997Marvaldi ( , 1998aMarvaldi ( , 1998bMarvaldi ( , 2003, the larva of Graptus possesses only this first character (4 vms). But 4 vms, observed constantly on larvae of all Entiminae, were also recorded on several other weevil taxa (e.g., Tychinii Gistel, 1848; Skuhrovec et al. 2015). On the other hand, the number of setae and shape of the larval body of G. triguttatus seem to be more similar to the Entiminae than to the Molytinae. But the fact that different larval stages and the pupae were found between the roots of dicotyledonous plants, outside of any plant tissues, supports the placement of Graptus (and Byrsopagini) in the subfamily Entiminae (Thompson 1992;Marvaldi 1997Marvaldi , 1998aAlonso-Zarazaga and Lyal 1999); even taking into account some small differences between the description of the mature larva of G. triguttatus presented by us and descriptions of L 1 presented by Van Emden (1952) and Marvaldi (1998a) (e.g., the lack of ts on L 1 versus lateral lobes of mature larva with a pair of ts). On the other hand, the rest of characters listed by Marvaldi (1998a) (e.g., 1 as; 4-6 pds on Abd. 8; sensillum cluster between mes 1 and mes 2 ; mandibles with bidentate apex; four anal lobes) have been mentioned in both descriptions.
From morphological data, the taxonomic position of Byrsopagini is apparently not so clear, but its placement within Entiminae is more plausible than in Molytinae; this is in line with the results of the cladistic and phylogenetic analyses performed by Marvaldi (1997Marvaldi ( , 1998a and Stüben et al. (2015).
Moreover, a detailed analysis of the structure of antennae disclosed only ostensible similarities between Graptus and Molytinae, e.g. Mitoplinthus caliginosus described by Sprick and Gosik (2014). First of all, the larva of M. caliginosus shows a larger number and variety of sensilla than the Graptus larva. The shape of the antennal sensorium is quite variable in both genera as well. Finally, the sensorium on Mitoplinthus is elongate and pointed, whereas it is stout and rounded in Graptus (Figs 71-76 a conical instead of a cushion-like sensorium in Graptus larvae is difficult to explain. Van Emden (1952) described the larva of Philopedon plagiatum with "broader and more rounded" sensoria. The same type of structure was found in larvae obtained during this study (Fig. 51). Van Emden (1950) did not recognize any kind of spiracles of the Graptus triguttatus larva, and finally marked them as "?". At first view they could be recognized as bicameral, but under high magnification (40×) they turned to appear annularly (Figs 77, 78).
The description of the pupa of Graptus triguttatus did not provide any relevant features that could confirm a replacement of this genus in the Entiminae. It seems to be especially difficult to distinguish them from other subfamilies due to the large diversity of both, shape and chaetotaxy, and the absence of features strictly characteristic of Entiminae pupae Sprick 2012a, 2013). The absence of urogomphi in pupae of G. triguttatus corresponds with pupae of some species of Otiorhynchus Germar, 1822 with strongly reduced urogomphi (Gosik and Sprick 2012a) and with the pupa of Liophloeus tessulatus (Müller, 1776), which does not bear urogomphi at all. Unfortunately, pupae of Molytinae are also characterized by variable development of urogomphi: they are well developed in the genera Pissodes Germar, 1817 and Trachodes Germar, 1824 (Scherf 1964) but almost absent in Homalinotus Schönherr, 1823 (De Oliveira Lira et al. 2017).

II. Some aspects of biology
Graptus triguttatus. In January and April 2013 many larvae of different instars were found, and most of them were already mature (Figs 85-87), which indicates hibernation in a higher instar and no development in this early time of the year. In May and June the proportion of mature larvae continued to grow, and pupae were also recorded. We conclude that teneral adults will be present from July onward. The biology is nevertheless poorly known. Dudich (1921) stated that adult weevils are mainly present from mid-April to mid-May. These findings are difficult to explain. If the species directly overwinters in soil after emerging from the pupa, the presence of winter larvae  of mainly higher instars is unexpected, and if the species appears on the soil surface in summer after reproduction, then the maximum of adults in April and May is hard to explain. It seems probable that there is a high degree of overlap between different generations; the life-cycle has still to be clarified.
Peritelus sphaeroides. According to Hoffmann (1963) and Dieckmann (1980), eggs are deposited in April and May and the new generation of adults emerges from June to August. Our data, with mature larvae from the end of August (Figs 83, 84), November and December (most larvae) until March of the following year, and two pupae found in December (Gosik and Sprick 2013), may suggest that there is a much longer period of larval development than given by the sources cited above. Probably there is a degree of overlap between the generations, as it was suggested for Graptus triguttatus and as it was often found in Otiorhynchus species (e.g., Gosik et al. 2016). Apparently, there is a need for regular search of larvae and pupae in the field to clarify the life-cycle.
Philopedon plagiatum. A description of the pupa is not available, even though the species attracted attention occasionally in beet fields, horticultural crops and pine plantations in sandy areas (Figs 79-82) (e.g., Dieckmann 1980;Brendler et al. 2008). Although teneral adults were found mainly in April (Dieckmann 1980), the period of pupation cannot be ascertained from this fact. In several species, such as Otiorhynchus raucus or O. singularis, pupation occurs in mid-summer and adults overwinter with smooth cuticula in their pupal chambers until next spring (Gosik et al. 2017). Morris (1987) found a young adult weevil with mandibular appendages in September. The fact of a long development and overlap of generations, as also suggested for the two species treated before, Graptus triguttatus and Peritelus sphaeroides, can be directly concluded from the data on P. plagiatum given by Dieckmann (1980). Morris' observations support these findings; he also stated that the species does not have a simple life-cycle, and he already supposed that it develops over two years. This seems to be probable and the best explanation for the data presented previously.