Distinction of Indian commercial lac insect lines of Kerria spp. (Homoptera: Coccoidea) based on their morphometrics.

The lac insects belong to the genus Kerria (Hemiptera: Coccoidea: Kerriidae) and are commercially exploited worldwide for the production of lac, which comes from their waxy test and has diverse industrial applications. The insects are maintained by the Indian Institute of Natural Resins and Gums as distinctive lines that are cultivated and commercialized in the lac producing areas of India. The lines are all considered to belong to the genus Kerria but without validation of their taxonomic characters, and their identity to species has not been ascertained. This study used single-factor analysis of variance and several multivariate analyses, such as principal component analysis, discriminant function analysis, and canonical discriminant analysis to explore the morphometrics of some of the adult female lac insect lines. The results have enabled the identification of some taxonomically significant characters in adult females, which has grouped the 32 lac insect lines studied into 15 species along with validation of the most significant characters. Distinctive grouping patterns for the species of Kerria have been brought out using morphometrics.

Scale insects (Sternorrhyncha: Coccoidea) are phytophagous insects found in all terrestrial zoogeographical regions except Antarctica, with ∼7,500 species in ∼30 families ( Ben-Dov et al. 2014 ). These are generally divided into two informal groups, the archaeococcoids and the neococcoids, based on the presence or absence of abdominal spiracles in the adult female. The neococcoids form a monophyletic group with 17 families and the Coccidae and the Tachardiidae form sister groups ( Cook et al. 2002 ). The family Tachardiidae (=Kerriidae), which includes lac insects, consists of nine genera and 100 species ( Ben-Dov et al. 2014 ). Lac insects (Hemiptera: Coccoidea: Tachardiidae) are morphologically distinctive scale insects that produce a gum-like or resinous secretion that forms a hard cover over the body ( Chamberlin 1923 , Varshney 1976 ). The word “lac” is derived from a Sanskrit word which mean “hundred thousand,” indicating the gregarious habit of this insect ( Krishnaswami 1962 ). These insects belong to the genus Kerria and the most commonly cultivated species is Kerria lacca (Kerr). The species of Kerria are distributed throughout India but occur as isolated patches in a variety of habitats ( Varshney 1976 , Ramani et al. 2007 ).

Lac insects yield three commercially important products: resin, dye, and wax, which have major applications in a wide range of industries ( Varshney 1976 , Ramani et al. 2007 ). These lac products are preferred over other products due to their unique properties along with their environmental safety ( Saha et al. 2011 ).

The commercially exploited species of lac insect belong to many distinctive genetic lines and these are maintained and cultivated by the Indian Institute of Natural Resins and Gums (IINRG). These genetic lines are commercially exploited for lac production in different parts of India. Taxonomy of the lac insect is based on the monograph and its supplement by Chamberlin ( 1923 , 1925 ) as well as subsequent works by Kapur (1958) , Varshney (1976) , and Kondo and Gullan (2007) . All these commercially cultivated lines have been placed in the genus Kerria but without validation of their taxonomic characters and thus each line is commercially cultivated without a proper identification. The diversity and cultivation complexities of these lines require a critical analysis through a study of their morphology and morphometrics. As there is much variability in their morphology, with significant overlapping of characters, these need to be analyzed and the most important characters clarified. Females are highly degenerative and undergo considerable changes in size and shape during sexual maturation, posing a challenge in their identification and so this intraspecific variation needs to be critically analyzed.

Hence, this study used single-factor analysis of variance (ANOVA) and multivariate analyses such as principal component analysis (PCA), discriminant function analysis (DFA), and canonical discriminant analysis (CDA) to explore the morphometrics of the commercial lac insect lines in India.

Materials and Methods

Collection and Preparation of Specimens

Thirty-two female lac insect lines were studied ( Table 1 ). These included species of Kerria from the principal lac growing states, geographical races, some inbred lines, and the infra-subspecific forms kusumi and rangeeni. Kusumi and rangeeni are two distinct forms of lac insects, the latter thriving on Butea monosperma (Fabaceae) but not on Schleichera oleosa (Sapindaceae), which is a preferred host of kusumi. The samples were obtained from the cultures maintained on potted Flemingia macrophylla (Fabaceae), a lac host plant kept under culture conditions at the Lac Insect Field Gene Bank of National Lac Insect Germplasm Center, IINRG campus, Ranchi (23° 19′51″ N and 85° 22′18″ E; elevation of 2,080 ft). These cultures are enclosed in synthetic mesh sleeves to exclude parasitoids and predators and are regularly sprayed with fungicide carbendazim (0.01%). In order to prepare the specimens for morphological studies, mature females were scraped from the twigs and placed in 100% ethyl alcohol for 48 hr to dissolve their resinous covering. The specimens were then cleaned carefully under a stereozoom microscope with a brush to remove any excess wax. These cleaned insects were preserved in 90% ethyl alcohol in a 1.5 ml eppendorf tube for further studies.

Table 1.

Lines of Kerria spp. studied and their species groups with locality and host

Sl. no.	Species groups	Line	Locality of collection	Host
1	Kerria brancheata (Varshney) group	LIK0014	Jammu, Jammu & Kashmir	Ziziphus mauritiana
		LIK0027	Silli, Jharkhand	Schleichera oleosa
		LIK0045	Experimental	Flemingia macrophylla
		LIK0064	Varanasi, Uttar Pradesh	Ficus religiosa
2	Kerria chamberlini (Varshney) group	LIK0015	Ambaji, Banaskantha, Gujarat	F. religiosa
		LIK0060	Purulia, West Bengal	Butea monosperma
		LIK0061	Bankura, West Bengal	Z. mauritiana
3	Kerria chinensis (Mahdihassan) group	LIK0023	Thailand	Albizia saman
4	Kerria dubeyi (Ahmad & Ramamurthy)	LIK0008	Bangalore, Karnataka	Ficus bengalensis
5	Kerria ebrachiata (Chamberlin) group	LIK0004	Palamau, Jharkhand	B. monosperma
		LIK0005	Bokaro, Jharkhand	B. monosperma
6	Kerria fici (Green) group	LIK0013	Ludhiana, Punjab	Z. mauritiana
7	Kerria indicola (Kapur) group	LIK0020	Echoda, Andhra Pradesh	Peltophorum ferrugineum
		LIK0029	Korba, Chhattisgarh	B. monosperma
		LIK0048	Experimental	F. macrophylla
8	Kerria lacca (Kerr) group	LIK0011	Udaipur, Rajasthan	F. religiosa
		LIK0012	Jhalod, Rajasthan	F. religiosa
		LIK0017	Ahmednagar, Maharashtra	Z. mauritiana
		LIK0018	Aurangabad, Maharashtra	P. ferrugineum
		LIK0028	Bokaro, Jharkhand	B. monosperma
		LIK0047	Experimental	F. macrophylla
		LIK0062	Medinipur, West Bengal	A. saman
9	Kerria maduraiensis (Ahmad & Ramamurthy)	K . maduraiensis	Madurai, Tamil Nadu	A. saman
10	Kerria manipurensis (Ahmad & Ramamurthy)	K . manipurensis	Churachandpur, Manipur	Malvaviscus penduliflorus
11	Kerria pennyae (Ahmad & Ramamurthy)	LIK0003	Sundergarh, Orissa	S. oleosa
12.	Kerria pusana (Misra) group	LIK0001	Korba, Chhattisgarh	S. oleosa
		LIK0039	Selection	S. oleosa
		LIK0040	Selection	S. oleosa
		LIK0065	Bankhedi, Madhya Pradesh	Original host not known
13	Kerria sharda (Mishra & Sushil) group	LIK0007	Sarat, Mayurbanj, Orissa	S. oleosa
14	Kerria thrissurensis (Ahmad & Ramamurthy)	LIK0010	Thrissur, Kerala	Amhertia nobilis
15	Kerria varshneyi (Ahmad & Ramamurthy)	LIK0063	Patiala, Punjab	Z. mauritiana

These alcohol-preserved specimens were slide mounted following the technique of Jena et al. (2011) . Briefly, the specimens were placed in 10% potassium hydroxide overnight to soften the internal tissue. They were then washed thoroughly in distilled water with 8–10 changes and then placed in 1% glacial acetic acid where a small incision was made on the lateral aspect of the body using a scalpel in order to remove the internal contents. The specimens were then cleaned thoroughly with fine needles and a brush and placed in polychromatic stain for about 20 min. They were then dehydrated through grades of ethyl alcohol of 70%, 90%, and 100% followed by clearing in 30%, 50%, and 80% xylene before preparing a permanent mount in Distrene, Plasticiser, Xylene (DPX). Finally, the slide mounts were dried on a hot plate at 45–60°C. These permanent microslides were made using Leica EZ4 stereozoom microscope.

Selection and Measurement of Characters

Using the morphological characters of adult female K. lacca taken from Chamberlin ( 1923 , 1925 ), Kapur ( 1958 , 1962 ), Varshney ( 1976 , 1985 ), Zhang (1993) , Mishra and Sushil (2000) , Lit and Gullan (2001) , Lit ( 2002a , b ), Kondo and Gullan (2007) , a total of 65 characters were identified for morphometric analyses. A standardization experiment using 30 specimens of each line, in all at a time just before these reach maturity, was undertaken to identify the characters which were most stable and consistent. These resulted in the selection of 50 characters, which had been supported by the single-factor ANOVA. These selected characters were measured and their morphology observed at magnifications between 100× and 1,000× using a Leica DM1000 phase contrast microscope with a micrometer eyepiece. The measurements are as in the slide-mounted specimens. The measurements of width used in the study are as follows: 1) width at apex—width taken at clypeolabral shield position, i.e., middle of tentorium; 2) width at middle—width taken where it is maximum, generally taken at the middle of the body; and 3) width at base—width taken at the position of base of anal tubercle.

Statistical Analysis

Univariate one-way single-factor ANOVA was performed individually for all the characters to select those that were significant as a prelude to identifying the potential characters ( Kalaisekar et al. 2012 ). These morphometrics were then analyzed using multivariate statistical approaches ( Tabachnick and Fidell 2006 ) as follows: PCA (SAS procedure, PRINCOMP, SAS version 9.1.3, SAS Institute Inc., Cary, NC), without any prior assumption of groupings, assesses the components for total variation among the specimens by calculating linear combinations of variables that explain the maximum of total variation. PCA was also used as a dimension-reducing technique. CDA (SAS procedure, CANDISC) calculates linear combinations of variables that maximize the separation of means of previously defined classes. Contribution of the variables best summarizing the differences between classes is revealed by this technique. Since DFA (SAS procedure, DISCRIM) maximizes the variation among groups, it was used to separate groups. DFA also determines the potential misclassification of specimens and assesses the utility of characters used.

These analyses were carried out in two batches: one with each of the 30 lines and in the other with seven species of Kerria , namely Kerriachinensis , Kerriamanipurensis , Kerriamaduraiensis , Kerriathrissurensis , Kerriapennyae , Kerriadubeyi , and Kerriavarshneyi ( Ahmad et al. 2013a , b ), to validate the species described. The sample size for each of the 30 lines was 30 and that for each of the seven species was 10.

Results and Discussion

Morphometrics and Species Distinctions

The 32 lac insect lines fell into two broad categories based on the structure of the anal tubercle, i.e., whether the tubercle is elongated or abbreviated. Both groups were then subdivided based on the shape and status of brachia into five groups: those with an elongated tubercle into three subgroups: 1) brachia elevated and cylindrical, 2) brachia elevated and club shaped, and 3) brachia sessile and club shaped; and those with an abbreviated tubercle into two subgroups: 1) brachia elevated and club shaped and 2) brachia sessile and club shaped, as shown in Fig. 1 . Based on the key to the adult females of Kerria , the species groups were differentiated.

Fig. 1.

Hierarchical flow diagram for the classification of 32 lines studied based on the characters of anal tubercle and brachia with distinct character for each species.

Morphometrics and Taxonomic Characters

The evaluation of some taxonomic characters using one-way ANOVA revealed that 50 were statistically significant ( P ≤ 0.01) ( Table 2 ). These characters were subjected to PCA analyses. The first five principal components (PCs) with an eigenvalue more than 1.0 accounted for 45.5% of the total variation ( Table 3 ). The first two PCs, i.e., PC1 and PC2, together explained about 25.6% of the total variation, with PC1 explaining 15.3% and PC2 explaining about 10.3%, respectively. These had positive loading for seven original variables, including the number of ducts in each marginal duct cluster (MDC), length of anal tubercle, length of pre-anal plate, distance of anterior spiracle from crater rim, length of brachia, length of pedicel, and total length of dorsal spine. The other PCs, i.e., PC3, PC4, and PC5, explained 8.5%, 6.7%, and 4.8% of the total variation, respectively. As the first two PCs accounted for 25.6% of the variability, those characters with maximum loadings were considered to be the major sources of variation. The plot for the first two PCs, i.e., PC1 and PC2, are shown in Fig. 2 , and the clusters emphasize the grouping of the lac insect lines. A compact clustering was observed for the lines LIK0023 ( K . chinensis ), LIK0010 ( K . thrissurensis ), LIK0001, LIK0039, LIK0040, and LIK0065 ( K . pusana group), LIK0008 ( K . dubeyi ), and LIK0003 ( K . pennyae ) in the first, second, and third quadrants, respectively, with the rest of the lines mostly overlapping.

Table 2.

Statistically significant characters ( P ≤ 0.01) for the morphometrics of 30 lines or species of Kerria

Sl. no.	Characters	Acronym
1	Length of canellar band	CL
2	Distance of anterior spiracle from crater rim	DCR
3	Length of brachia	BrL
4	Number of ducts in MDC III	MDCIII
5	Number of ducts in MDC II	MDCII
6	Number of dimple on brachia II	DII
7	Number of dimple on brachia I	DI
8	Number of ducts in MDC V	MDCV
9	Number of ducts in MDC I	MDCI
10	Length of pedicle	PL
11	Total length of dorsal spine	TDSL
12	Number of ducts in MDC IV	MDCIV
13	Number of ducts in MDC VI	MDCVI
14	Diameter of brachial plate	BPD
15	Width of crater	CW
16	Width of anterior spiracle	ASW
17	Width of supra-anal plate	SPW
18	Length of antennae	AL
19	Length of anal tubercle	ATL
20	Length supra-anal plate	SPL
21	Number of spiracular pores	NSP
22	Perivulvar pore cluster II	PVCII
23	Width of body at middle	BWM
24	Number of antennal segments	NAS
25	Width of pre-anal plate	PAW
26	Pedicle width at apex	PeWA
27	Number of antennal setae	NASe
28	Length of pre-anal plate	PAL
29	Perivulvar pore cluster I	PVCI
30	Length of antennal segment III	ALIII
31	Body length	BL
32	Length of spine	SL
33	Pedicle width at base	PeWB
34	Length of posterior spiracle	PSL
35	Length of anterior spiracle	ASL
36	Number of star pores	NSPo
37	Width of body at base	BWB
38	Width of body at apex	BWA
39	Length of oral lobe	OLL
40	Length of antennal segment IV	ALIV
41	Length of anal fringe	FL
42	Length of antennal segment II	ALII
43	Width of clypeolabral shield	WCS
44	Length of clypeolabral shield	LCS
45	Perivulvar pore cluster opening V	PoCV
46	Perivulvar pore cluster opening III	PoCIII
47	Perivulvar pore cluster opening II	PoCII
48	Perivulvar pore cluster opening I	PoCI
49	Length of antennal segment I	ALI
50	Perivulvar pore cluster opening IV	PoCIV

Table 3.

Proportion of variation and variable coefficients of the first five PCs for PCA and total sample standardized canonical coefficients of CDA for the 30 lines of Kerria spp.

	Principal component 1	Principal component 2	Principal component 3	Principal component 4	Principal component 5	Canonical axis 1	Canonical axis 2
CL	0.149	0.158	−0.104	−0.297	−0.010	1.420	0.520
DCR	−0.159	0.292	−0.058	0.006	−0.011	−0.017	0.747
BrL	−0.153	0.265	−0.078	0.043	−0.023	−0.111	0.246
MDCIII	0.319	0.070	−0.048	−0.005	−0.002	0.497	0.026
MDCII	0.314	0.045	−0.055	0.000	0.003	0.232	−0.233
DII	−0.102	−0.156	0.061	0.342	0.035	−0.290	−0.199
DI	−0.114	−0.153	0.049	0.339	0.034	−0.374	−0.033
MDCV	0.310	0.049	−0.077	−0.014	−0.024	0.147	−0.099
MDCI	0.309	0.033	−0.057	−0.008	−0.002	−0.054	−0.203
PL	−0.154	0.238	−0.166	0.102	0.016	0.188	0.578
TDSL	−0.163	0.263	−0.167	0.138	0.020	−0.049	0.252
MDCIV	0.317	0.063	−0.045	−0.003	−0.017	0.244	−0.222
MDCVI	0.302	0.045	−0.040	−0.025	0.002	0.034	0.013
BPD	−0.087	−0.107	0.068	0.277	0.022	−0.684	0.097
CW	−0.147	0.094	−0.094	0.096	0.030	0.106	0.439
ASW	0.143	0.201	−0.100	0.085	−0.028	0.157	−0.212
SPW	0.146	0.048	0.033	0.313	−0.040	0.059	−0.722
AL	0.189	−0.125	0.066	0.192	−0.022	−0.093	−0.358
ATL	−0.008	0.353	0.004	0.083	−0.024	−0.867	0.597
SPL	0.020	0.194	−0.047	0.279	−0.051	0.299	0.042
NSP	0.087	0.114	−0.112	−0.008	−0.041	0.102	0.022
PVCII	0.133	0.067	−0.032	0.190	0.034	−0.051	−0.009
BWM	0.035	0.173	0.429	−0.008	0.087	−0.574	−0.800
NAS	0.110	−0.099	0.127	0.044	0.060	−0.051	−0.173
PAW	0.115	0.106	0.001	0.261	−0.006	−0.143	−0.196
PeWA	−0.068	0.013	−0.007	0.171	0.009	−0.236	0.289
NASe	−0.097	0.048	−0.058	−0.067	−0.022	−0.002	0.126
PAL	−0.021	0.323	0.034	−0.056	0.000	0.925	−0.257
PVCI	0.100	0.046	−0.048	0.188	0.023	0.068	0.079
ALIII	0.162	−0.128	0.056	0.115	0.026	−0.039	0.055
BL	0.024	0.196	0.417	−0.015	0.073	0.091	0.081
SL	−0.111	0.193	−0.098	0.135	0.026	0.004	−0.082
PeWB	0.032	−0.109	0.025	0.226	0.077	−0.012	−0.045
PSL	0.107	0.029	−0.062	0.148	−0.032	−0.062	−0.061
ASL	0.061	0.173	−0.105	0.067	−0.092	0.068	−0.020
NSPo	0.069	0.018	−0.062	−0.088	−0.011	−0.001	0.012
BWB	0.036	0.180	0.418	−0.002	0.094	0.233	0.269
BWA	0.037	0.169	0.427	−0.012	0.081	0.125	0.298
OLL	0.112	−0.021	−0.030	0.101	0.000	−0.093	−0.203
ALIV	0.025	−0.026	0.100	0.036	0.057	0.009	0.168
FL	−0.032	0.057	−0.035	−0.085	−0.020	0.136	0.048
ALII	0.052	−0.033	−0.039	0.113	−0.094	0.181	0.031
WCS	0.006	−0.003	0.032	−0.025	−0.022	0.027	−0.034
LCS	−0.011	0.023	0.187	0.000	−0.324	0.002	−0.091
PoCV	0.006	0.003	−0.124	0.028	0.466	−0.024	−0.190
PoCIII	0.022	0.056	−0.062	−0.004	−0.049	0.107	−0.021
PoCII	0.043	0.064	−0.066	0.066	−0.037	0.003	0.001
PoCI	0.055	0.025	−0.054	−0.004	−0.107	−0.011	−0.129
ALI	−0.011	−0.014	0.063	0.050	−0.562	0.018	−0.156
PoCIV	0.050	0.030	−0.073	−0.012	0.519	0.077	−0.005
Eigenvalues	7.629	5.158	4.236	3.347	2.389
Proportion of variation	15.3%	10.3%	8.5%	6.7%	4.8%

Fig. 2.

Scatter plot of PCs 1 and 2 showing the grouping of 30 lines of Kerria spp. Encircled regions showing the compact clustering for Kerria chinensis (LIK0023), Kerria pusana group (LIK0001, LIK0039, LIK0040, and LIK0065), Kerria pennyae (LIK0003), and Kerria dubeyi (LIK0008) in the first, second, and third quadrants, respectively.

CDA was carried out with priori grouping and using the lines as classification variables. The statistics used to test differences between the lines, namely Wilks’ λ, Pillai’s trace, Hotelling–Lawley Trace, and Roy’s greatest root, were found to be significant at P  < 0.0001. These statistics clearly show the significant contribution toward the model, with a lower Wilks’ λ (2.8 × 10 −7 ), holding true for all other statistics ( Table 4 ). The first two canonical correlations (89.8% and 88.9%) were very high, signifying their importance ( Table 5 ). The projection of the lines onto the first two canonical discriminant axes is shown in Fig. 3 . The analysis was able to extract differences between the lines LIK0001, LIK0039, LIK0040, and LIK0065 ( K . pusana group), LIK0003 ( K . pennyae ), LIK0008 ( K . dubeyi ), and LIK0023 ( K . chinensis ), but there was extreme overlapping among the rest of these lines. The first canonical root clearly discriminated LIK0023 ( K . chinensis ) from the rest, with the main contribution being from canellar band length, while the second canonical root was not particularly helpful in discriminating between any lines ( Table 3 ). This clustering obtained from CDA confirmed the grouping brought out by PCA. A cross-validation of group membership was performed identifying the misclassification of specimens and assessing the utility of the selected measurements/observations used. Overall, 78% of the classifications were correctly attributed to species, with relatively few (22%) misclassifications ( Table 6 ). The results of cross-validation accurately identified 100% of specimens to the lines LIK0003 ( K . penny ae), LIK0008 ( K . dubeyi ), LIK0017 ( K . lacca group), and LIK0023 ( K . chinensis ); >90% to the lines LIK0005 ( Kerriaebrachiata group), LIK0007 ( Kerriasharda ), LIK0040 ( K . pusana group), LIK0045 ( Kerriabrancheata group), LIK0047 ( K . lacca group), and LIK0065 ( K . pusana group); and >80% to the lines LIK0001 ( K . pusana group), LIK0012 ( K . lacca group), and LIK0063 ( K.varshneyi ), respectively. Thus, the DFA results helped to identify these 13 species groups based on the merit of each of the morphological characters used in the analyses.

Table 4.

Multivariate statistics and F approximations for the 30 lines of Kerria spp.

Statistic	Value	F value	Num DF	Den DF	Pr >  F
Wilks’ λ	2.80 × 10 −7	11.83	1,450	20,841	<0.0001
Pillai’s trace	9.2648	7.97	1,450	24,621	<0.0001
Hotelling–Lawley trace	33.7649	19.07	1,450	14,529	<0.0001
Roy’s greatest root	8.8518	150.3	50	849	<0.0001

Table 5.

Canonical correlation analysis for the 30 lines of Kerria spp.

	Canonical correlation	Adjusted canonical correlation	Approximate standard error	Squared canonical correlation
1	0.948		0.003	0.898
2	0.943		0.004	0.889
3	0.881	0.870	0.007	0.777
4	0.849	0.835	0.009	0.721
5	0.793	0.768	0.012	0.629
6	0.777	0.757	0.013	0.603
7	0.749	0.727	0.015	0.561
8	0.711	0.684	0.017	0.505
9	0.669	0.632	0.018	0.447
10	0.646	0.616	0.019	0.417
11	0.615	0.581	0.021	0.379
12	0.585	0.545	0.022	0.342
13	0.565	0.537	0.023	0.319
14	0.537	0.510	0.024	0.288
15	0.473		0.026	0.224
16	0.465		0.026	0.216
17	0.416		0.028	0.173
18	0.404		0.028	0.163
19	0.355		0.029	0.126
20	0.334		0.030	0.112
21	0.313		0.030	0.098
22	0.303		0.030	0.092
23	0.281		0.031	0.079
24	0.242		0.031	0.059
25	0.206		0.032	0.042
26	0.184		0.032	0.034
27	0.183		0.032	0.033
28	0.141		0.033	0.020
29	0.128		0.033	0.016

Fig. 3.

Scatter plot of the results of CDA for the 30 lines of Kerria spp. showing a similar clustering for lines as in PCA.

Table 6.

Classification matrix of the DFA for 30 lines of Kerria spp. studied where rows = observed classification and columns = predicted classification

Lines	% Age	01	03	04	05	07	08	10	11	12	13	14	15	17	18	20	23	27	28	29	39	40	45	47	48	60	61	62	63	64	65
01	87	26	0	0	0	0	2	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	1	0	0	0	0
03	100	0	30	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
04	57	0	0	17	1	0	0	0	0	0	0	0	0	0	2	0	0	0	0	0	0	0	0	0	0	3	2	0	0	5	0
05	93	0	0	0	28	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	2	0	0	0	0	0
07	97	0	0	0	0	29	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
08	100	0	0	0	0	0	30	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
10	77	0	0	0	0	0	0	23	0	0	0	1	0	0	0	1	0	0	2	0	0	0	1	0	0	0	2	0	0	0	0
11	70	0	0	1	0	0	0	0	21	2	0	0	2	0	0	0	0	0	0	1	0	0	0	0	0	0	1	0	0	2	0
12	80	0	0	0	0	0	0	0	4	24	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	2	0	0
13	73	0	0	0	0	0	0	0	0	0	22	2	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0	0	1	3	0
14	67	0	0	2	0	1	0	0	0	0	3	20	0	0	0	1	0	0	1	1	0	0	0	1	0	0	0	0	0	0	0
15	77	0	0	0	1	0	0	0	0	2	0	0	23	0	1	0	0	0	1	1	0	0	0	0	0	0	0	0	0	1	0
17	100	0	0	0	0	0	0	0	0	0	0	0	0	30	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
18	67	0	0	1	0	0	0	0	0	0	1	0	4	0	20	0	0	0	0	0	0	0	0	0	0	1	0	0	0	3	0
20	70	0	1	2	0	0	0	0	0	0	0	0	0	0	1	21	0	0	0	2	0	0	0	1	0	2	0	0	0	0	0
23	100	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	30	0	0	0	0	0	0	0	0	0	0	0	0	0	0
27	57	0	0	0	0	0	0	1	0	0	0	0	1	0	1	0	0	17	4	1	0	0	0	0	1	0	1	0	0	3	0
28	63	0	0	2	2	0	0	0	1	0	0	0	1	0	0	0	0	1	19	0	0	0	1	0	0	3	0	0	0	0	0
29	57	0	0	0	0	0	0	0	2	0	0	1	1	1	0	6	0	0	0	17	0	0	0	0	0	2	0	0	0	0	0
39	63	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	19	4	0	0	0	1	0	1	0	0	4
40	97	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	29	0	0	0	0	0	0	0	0	1
45	90	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	2	0	0	0	27	0	1	0	0	0	0	0	0
47	90	0	0	0	0	0	0	0	1	0	0	0	0	1	0	0	0	0	0	0	0	0	0	27	0	0	0	0	0	1	0
48	63	0	0	1	0	0	0	0	7	0	0	0	0	0	0	0	0	0	0	0	0	0	0	1	19	1	0	0	0	1	0
60	47	0	0	2	1	0	0	0	1	0	0	0	1	0	0	0	0	0	1	0	0	0	0	0	0	14	1	6	1	2	0
61	73	0	0	2	0	0	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	2	22	3	0	0	0
62	77	0	0	0	1	0	0	0	1	0	0	0	0	0	0	0	0	0	1	0	0	0	0	0	0	2	2	23	0	0	0
63	80	0	0	1	0	0	0	0	0	0	4	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	24	1	0
64	70	0	0	1	0	0	0	0	1	0	1	0	0	0	2	0	0	0	0	0	0	0	0	0	0	2	0	0	2	21	0
65	90	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	2	0	0	0	0	0	0	0	0	27
Total	78	27	31	33	34	30	33	24	39	28	31	24	33	32	27	29	30	20	31	23	19	35	29	31	21	36	32	33	30	43	32

Establishment of Species Classification of New Species Described

The status of recently described six new species— K . manipurensis (Ahmad & Ramamurthy), K . thrissurensis (Ahmad & Ramamurthy), K . maduraiensis (Ahmad & Ramamurthy), K . pennyae (Ahmad & Ramamurthy), K . dubeyi (Ahmad & Ramamurthy), and K . varshneyi (Ahmad & Ramamurthy) ( Ahmad et al. 2013a , b )—and K . chinensis (Mahdihassan) were supported by the multivariate analyses (PCA, CDA, and DFA). The PCA indicated that the first 10 PCs with eigenvalues more than 1 accounted for 79.9% of the total variation. Contribution of variables to the first three PCs accounted for 51.3% of the total variation ( Table 7 ). PC1 reflected a generalized increase in the values of five characters: distance of anterior spiracle from crater rim, number of ducts in each marginal duct cluster, width of anterior spiracle, length of anal tubercle and pre-anal plate length with a decrease in only one character, and number of dimples on brachial plate. The main contributions to PC2 were from five characters: brachial plate diameter, crater width, body width, width of anterior spiracle and width of supra-anal plate and to PC3 were from three characters: total length of dorsal spine, length of spine, and length of anal fringe. The other PCs, namely PC4–PC10, explained 6.7%, 5.6%, 4.6%, 3.8%, 3.1%, 2.6%, and 2.2% of the total variation respectively, and therefore made little contribution toward explaining the variation. The differences in distribution across the common component of variation for the seven species of Kerria are evident in Fig. 4 . The results of PCA show the distinctiveness of the species studied except for a slight overlap between K . dubeyi and K . pennyae and for K . varshneyi and K . maduraiensis . A dispersed clustering was observed for both K . maduraiensis and K . varshneyi .

Table 7.

Proportion of variation and variable coefficients for the first three PCs of PCA and total sample standardized canonical coefficients of CDA for seven Kerria species

Variables	Component 1	Component 2	Component 3	Canonical axis 1	Canonical axis 2
CL	0.18	−0.114	−0.185	2.062	−1.804
DCR	0.204	0.099	0.147	−0.935	2.491
BrL	0.18	0.166	−0.006	1.400	0.129
MDCIII	0.222	−0.159	0.096	0.467	−2.020
MDCII	0.225	−0.145	0.077	0.199	−1.312
DII	−0.196	0.194	0.069	0.024	1.791
DI	−0.205	0.176	0.064	−0.835	2.858
MDCV	0.216	−0.171	0.048	0.917	1.414
MDCI	0.202	−0.185	0.087	−2.717	1.594
PL	0.052	0.131	0.324	−0.189	−2.352
TDSL	0.038	0.089	0.36	1.059	1.174
MDCIV	0.222	−0.169	0.037	2.952	2.138
MDCVI	0.213	−0.187	0.063	2.514	−0.113
BPD	−0.119	0.264	−0.122	−10.413	−2.389
CW	0.034	0.236	0.119	5.573	2.633
ASW	0.227	0.025	0.151	1.586	0.498
SPW	0.11	0.215	−0.085	2.098	−1.587
AL	0.117	0.174	−0.038	1.516	1.314
ATL	0.213	0.145	0.071	−0.556	0.780
SPL	0.156	0.199	0.171	0.154	1.859
NSP	0.147	−0.12	−0.146	−1.317	−0.032
PVCII	0.057	−0.153	0.164	−1.033	−0.117
BWM	0.158	0.23	−0.12	−2.905	−0.686
NAS	−0.005	−0.008	−0.065	0.123	−0.479
PAW	0.158	0.177	−0.177	−2.957	−0.899
PeWA	−0.068	0.132	0.055	0.540	0.424
NASe	−0.137	−0.034	0.149	0.177	−0.443
PAL	0.207	0.086	−0.001	0.000	0.000
PVCI	0.01	−0.088	0.16	1.152	1.229
ALIII	0.091	0.114	−0.076	−1.332	−0.189
BL	0.187	0.167	−0.164	−0.187	0.528
SL	−0.002	−0.016	0.266	0.000	0.000
PeWB	−0.164	0.047	−0.114	1.775	0.823
PSL	0.103	0.104	−0.194	0.642	−0.242
ASL	0.185	0.007	0.073	−1.085	−0.418
NSPo	−0.01	−0.157	−0.16	0.571	−0.314
BWB	0.166	0.18	−0.161	2.604	−0.178
BWA	0.157	0.238	−0.035	−0.826	−1.711
OLL	0.063	−0.002	−0.122	0.293	0.069
FL	0.079	0.027	0.308	−0.631	−0.337
ALII	0.078	0.102	0.066	0.762	0.674
WCS	−0.032	0.158	0.111	0.905	0.288
LCS	0.019	−0.156	−0.132	−0.705	−0.236
PoCV	−0.032	0.117	0.14	−0.354	0.131
PoCIII	0.074	0.077	−0.114	−0.349	−0.142
PoCII	0.08	0.019	−0.059	0.171	−1.020
PoCI	0.135	0.008	−0.128	0.223	−0.118
ALI	0.084	−0.089	−0.135	0.644	−0.475
PoCIV	0.026	0.085	0	0.714	−1.516
Eigenvalues	13.232	7.639	4.267
Proportion of variations	27.00%	15.60%	8.70%

Fig. 4.

Scatter plot of PCs 1 and 2 along the two axes for seven Kerria species with compact clustering for all except Kerria maduraiensis and little overlaps between K . dubeyi and K . pennyae and Kerria varshneyi with those of K . maduraiensis and Kerria thrissurensis . Symbols indicate species.

The CDA showed a highly significant Wilks’ λ value (1.0 × 10 −8 ), Pillai’s trace, Hotelling–Lawley Trace, and Roy’s greatest root ( P  < 0.0001) ( Table 8 ). The first two canonical correlations with squared canonical values 99.4% and 98.3% in canonical correlation analysis ( Table 9 ) were high, indicating their importance. Table 10 shows that the mean canonical variables with canonical roots having higher values for their respective variables (species) in canonical root 1 was able to separate the seven species studied, whereas canonical root 2 particularly separated K . chinensis and K . thrissurensis . The character brachial plate diameter contributed maximum (−10.413) to canonical root 1, toward the separation of species ( Table 7 ). The projection of species onto the first two canonical axes is shown in Fig. 5 . In the scatter plot, all the species were well separated with a compact clustering for each, except for a small overlapping between K . pennyae and K . varshneyi . This clustering obtained confirms the groups presented in the PCA without the overlaps.

Table 8.

Multivariate statistics and F approximations for seven Kerria species

Statistic	Value	F value	Num DF	Den DF	Pr >  F
Wilks’ λ	1.0 × 10 −8	9.17	282	110.24	<0.0001
Pillai’s trace	5.6189	6.9	282	132	<0.0001
Hotelling–Lawley trace	263.8778	14.54	282	57.996	<0.0001
Roy’s greatest root	158.1423	74.02	47	22	<0.0001

Table 9.

Canonical correlation analysis for seven Kerria species

	Canonical correlation	Adjusted canonical correlation	Approximate standard error	Squared canonical correlation
1	0.997		0.001	0.994
2	0.992		0.002	0.983
3	0.974	0.955	0.006	0.949
4	0.962	0.936	0.009	0.925
5	0.942		0.014	0.887
6	0.938		0.014	0.880

Table 10.

Mean canonical variables based on discriminant functions of the morphological characters for seven Kerria species

	Canonical root 1	Canonical root 2	Canonical root 3	Canonical root 4	Canonical root 5	Canonical root 6
K . varshneyi	−12.482	−2.908	−4.425	5.109	1.998	2.117
K . dubeyi	−13.940	0.915	5.408	1.991	−1.639	−3.775
K . pennyae	−9.432	−2.876	2.539	−5.887	3.062	1.616
K . chinensis	18.140	−12.964	2.420	1.296	−0.526	0.227
K . maduraiensis	−2.519	2.262	−2.574	−2.138	−5.451	2.362
K . manipurensis	7.186	2.206	−6.424	−2.033	1.122	−4.113
K . thrissurensis	13.046	13.364	3.057	1.663	1.434	1.566

Fig. 5.

Scatter plot for the results of CDA for seven Kerria species showing compact clustering of the species studied as in the PCA, with slightest of overlap between K . varshneyi and K . pennyae .

A validation analysis through DFA of group participation/composition was performed for the seven species under study and it was observed that 87% of the classification was correctly attributed ( Table 11 ). Also, the result of the validation analysis (DFA) correctly identified 100% of specimens to K . pennyae ; 90% to K . dubeyi , K . manipurensis , K . thrissurensis , and K . chinensis ; and 70–80% to K . varshneyi and K . maduraiensis , respectively.

Table 11.

Classification matrix of the DFA for seven Kerria species, where rows = observed classification and columns = predicted classification

	Percentage	K. manipurensis	K. maduraiensis	K. thrissurensis	K . pennyae	K . dubeyi	K. chinensis	K . varshneyi
K. manipurensis	90.0	9.0	1.0	0.0	0.0	0.0	0.0	0.0
K. maduraiensis	80.0	0.0	8.0	0.0	2.0	0.0	0.0	0.0
K. thrissurensis	90.0	0.0	1.0	9.0	0.0	0.0	0.0	0.0
K . pennyae	100.0	0.0	0.0	0.0	10.0	0.0	0.0	0.0
K . dubeyi	90.0	0.0	1.0	0.0	0.0	9.0	0.0	0.0
K. chinensis	90.0	1.0	0.0	0.0	0.0	0.0	9.0	0.0
K . varshneyi	70.0	0.0	1.0	0.0	1.0	1.0	0.0	7.0
Total	87.1	10.0	12.0	9.0	13.0	10.0	9.0	7.0

Taxonomic Characters and Their Validation

The results of these analyses revealed that there are 14 characters which are consistent, without significant intraspecific variations, and which helped to separate the lac insect lines into species and groups. Most of these characters were in agreement with the 11 major characters noted by earlier taxonomists ( Table 12 ). In this study, many characters have been added such as body widths (apex, middle, and base), number of star pores near the mouthparts, width of anterior spiracle, length of pre-anal plate (membranous extension below the supra-anal plate), length and width of supra-anal plate, pedicel length, spine length, width of pedicel at base, pedicel width at apex, total length of dorsal spine, perivulvar pore cluster openings, and length of antennal segments. These additional characters were not used by earlier taxonomists but were found to be significant in species delineation in our studies, while other characters (length of antennal segments, spine length, pedicel width at apex, number of star pores near the mouthparts, perivulvar pore cluster openings) did not separate species in our studies due to their high intraspecific variation and low character loading, as revealed in both univariate and multivariate analyses.

Table 12.

Taxonomic characters versus lac insect species delineations, new characters (*) with statistical significance

Taxonomic characters hitherto used	Taxonomic characters used now	Additional characters evaluated
Body width	Body width at middle	Body width at middle*
	Body width at apex	Body width at apex*
	Body width at base
Antennae	Length of antennal segments
Number of ducts in marginal duct cluster	Number of ducts in each marginal duct cluster	Length of pre-anal plate*
Length of anal tubercle	Length of anal tubercle	Width of supra-anal plate*
	Length of pre-anal plate	Width of anterior spiracle*
	Width of supra-anal plate	Length of antennal segments
	Length of supra-anal plate
Length of brachia	Length of brachia	Spine length
Brachial plate width	Brachial plate diameter	Number of star pores near mouthparts
Crater width	Crater width	Number of openings in perivulver pore clusters
Number of dimples on brachial plate	Number of dimples on brachial plate
Distance of anterior spiracle from crater rim	Distance of anterior spiracle from crater rim
Pedicel length	Pedicel length
	Spine length
	Width of pedicel at base
	Pedicel width at apex
Total length of dorsal spine	Total length of dorsal spine
Length of anterior spiracle	Width of anterior spiracle

Note. If should be noted that character states concerning membranous structures should be used with care because age may affect their size, e.g., apex, middle, and basal body width. In this study, cultured specimens were used, which tend to be very uniform in size; however, specimens found in the wild may not be as uniform. Insect body size in the field is going to depend on host plant vigor, so that, even if the lac insect population may be restricted to a single species of host plant, those plants could be either growing slowly due to poor conditions (in which case the lac insects will be small) or growing very vigorously, in which case the scales may grow larger. This could affect the repeatability of these measurements at the species level. Furthermore, in the field, for these measurements to be useful, adult females must be collected exactly at the right stage (just prior to the appearance of the crawlers).

The clustering of lines, as revealed through the PCA and CDA analyses, were identical for each species group, indicating the validity of the characters used. The PCA and CDA analyses produced a clear separation of lines and species, indicating that the differences were genetic rather than environmentally induced ( Padi and Hollander 1996 ). The genetic diversity of these lines shown through Random Amplified Polymorphic DNA profiling ( Ranjan et al. 2011 ) and Inter Simple Sequence Repeat Markers ( Saha et al. 2011 ) was found to be in agreement with the results of this study. The DFA with higher classification (78% and 87%) values supported the division of the lac insect lines into species groups, and has also helped in establishing their consistency. This study also provides an insight into the validity of the taxonomic characters deployed in the genus Kerria for the species delineation.