Appendix 1

Supplementary material for ‘Methods’ section

  1. Study area

There was considerable inter-annual variability in rainfall from January 2008 to December 2010. For the duration of the study from April 2008 to February 2010, the rainfall received during the monsoon season also varied. In 2008, most of the rainfall was received between June and November while in 2009, most of the rainfall was received between July and August. December to February for the duration of the study were characterised by < 100 mm rainfall.

  1. Study Design
  1. Quantification of lantana categories in the MFDP:

Lantana is quantified at the Mudumalai Forest 1. Dynamics Plot in terms of the area covered by the plant in an area of 10m x 10m. Lantana has been categorised as being ‘absent’, ‘present’, ‘common’, ’dense’ and ‘very dense’. In 2011, a field study was conducted in order to determine the biomass of lantana in each of these categories (Suresh et al, unpublished data). Ten 1m x 1m plots were set outside the MFDP in areas that were visually determined as belonging to one of the five lantana density categories. Lantana was harvested, oven-dried and weighed again to obtain dry weight biomass. While the lantana ‘absent’ category corresponded to no lantana, the biomass of the other categories were as follows: ‘present’ – 48.3 ± 15.6 g/m2, ‘common’- 503 ± 30.5 g/m2, ‘dense’ – 858.5 ± 154.2 g/m2 and ‘very dense’- 2038.5 ± 226 6 g/m2.

  1. List of species of woody plants seedlings which were tagged and monitored from April 2008 to March 2010. Sample size is the total number of individuals pooled for lantana absent (LA) and lantana dense (LD) plots. Repeated measurements on seedling heights were made once every two months.

Species / Family / Forest habitat preference / Habit / Sample size (LA plots) / Sample size (LD plots)
Albizia odoratissima (L.f.) Benth. / Fabaceae / Moist / Canopy tree / 0 / 3
Anogeissus latifolia (DC.)Wall. ex Guill. & Perr. / Combretaceae / Dry / Canopy tree / 15 / 20
Antidesma diandrum (Roxb.) Roth / Euphorbiaceae / Moist / Understory tree / 11 / 3
Ardesia solanacea Roxb. / Myrsinaceae / Moist / Shrub / 1 / 0
Bauhinia malabarica Roxb. / Fabaceae / Moist / Canopy tree / 1 / 0
Bridelia retusa (L.) Sprengel / Euphorbiaceae / Dry / Canopy tree / 1 / 0
Canthium dicoccum (Gaertner) Teijsm. & Binnend. / Rubiaceae / Dry / Understory tree / 31 / 10
Casearia esculenta Roxb. / Flacourtiaceae / Moist / Canopy tree / 7 / 1
Cassia fistula L. / Fabaceae / Ubiquitous / Understory tree / 14 / 7
Dalbergia latifolia Roxb. / Fabaceae / Moist / Canopy tree / 18 / 11
Dalbergia lanceolaria L.f. / Fabaceae / Dry / Understory tree / 13 / 12
Diospyros montana Roxb. / Ebenaceae / Dry / Canopy tree / 50 / 59
Flacourtia indica (Burman) Merr. / Flacourtiaceae / Ubiquitous / Understory tree / 10 / 0
Garuga pinnata Roxb. / Burseraceae / Ubiquitous / Canopy tree / 1 / 0
Grewia orbiculata Rottl. / Tiliaceae / Dry / Shrub / 4 / 7
Grewia tiliifolia Vahl / Tiliaceae / Dry / Canopy tree / 53 / 50
Helicteres isora L. / Sterculiaceae / Moist / Shrub / 21 / 45
Hymenodictyon orixense (Roxb.) Mabb. / Rubiaceae / Moist / Canopy tree / 1 / 1
Kydia calycina Roxb. / Malvaceae / Moist / Understory tree / 21 / 1
Lagerstroemia microcarpa Wt. / Lythraceae / Moist / Canopy tree / 2 / 2
Lannea coromandelica (Houtt.) Merr. / Anacardiaceae / Moist / Canopy tree / 1 / 0
Mitragyna parviflora (Roxb.) Korth. / Rubiaceae / Dry / Canopy tree / 0 / 1
Naringi crenulata (Roxb.) Nicolson / Rutaceae / Dry / Understory tree / 0 / 1
Phyllanthus emblica L. / Phyllanthaceae / Dry / Understory tree / 25 / 5
Pterocarpus marsupium Roxb. / Fabaceae / Ubiquitous / Canopy tree / 4 / 0
Catunaregam spinosa (Thunb.) Tirv. / Rubiaceae / Dry / Understory tree / 161 / 196
Radermachera xylocarpa (Roxb.) Schum. / Bignoniaceae / Dry / Canopy tree / 0 / 1
Schleichera oleosa (Lour.) Oken / Sapindaceae / Moist / Canopy tree / 29 / 30
Stereospermum personatum(Hassk.) Chatterjee / Bignoniaceae / Moist / Canopy tree / 0 / 1
Syzygium cumini (L.) Skeels / Myrtaceae / Moist / Canopy tree / 63 / 73
Tectona grandis L.f. / Verbenaceae / Ubiquitous / Canopy tree / 3 / 2
Terminalia crenulata / Combretaceae / Dry / Canopy tree / 1 / 0
Wrightia tinctoria (Roxb.) R.Br. / Apocynaceae / Dry / Understory tree / 1 / 0
  1. Rationale for using height as a measure of seedling growth:

Owing to the deciduous nature of many of the species examined, we used height instead of leaf area or leaf number to quantify seedling growth response. Above-ground tissue was also frequently lost mostly due to desiccation in the dry season and rarely due to herbivory by large mammals at other times, thus stem diameter could not be used as reliable measures of growth.

  1. Statistical Analyses
  1. Mixed effects models specified for community and species level analyses -

Growth Interval / Fixed Effects / Random Effects
All species (community model) / 2 & 4 months / Lantana Density
Inter-census Rainfall
Habitat preference
All 2 way interactions / Plot ID
All species (community model) / 22 months / Lantana Density
Habitat preference
All 2 way interactions / Plot ID
Most Abundant species (Catunaregam spinosa, Syzygium cumini, Grewia tiliifolia, Diospyros montana) / 2 & 4 Months / Lantana Density
Inter-census rainfall
All 2 way interactions / Plot ID
Most Abundant species (Catunaregam spinosa, Syzygium cumini, Grewia tiliifolia, Diospyros montana) / 22 Months / Lantana Density / Plot ID

Rainfall was not considered as a fixed effect in analysing 22-month growth rate as the variation in this predictor was very small for this duration. Habitat preference was not considered in species-specific analyses because each species belonged to single habitat preference guild.

  1. Use of 95% CI for determining significance of model terms :

For determining differences between the levels of the categorical variables – lantana density and species’ habitat preference, we used default orthogonal contrasts specified in R, with all levels being compared to the alphabetically first level. The parameter estimates in an R mixed effects model output correspond to differences between means of all levels of a categorical variable from the alphabetically first reference level. For example, in the default ‘treatment contrast’ of linear mixed model packages, the mean growth rates of moist- forest preferring and ubiquitous species would be compared with the mean growth rate of dry forest species. If the 95% CI of an estimate overlapped zero (say, the lower CI is a negative number, while the upper CI is a positive number), one cannot say with certainty if the mean of one level is greater or lesser than the reference level. Estimates with confidence intervals that did not overlap with zero could be conclusively inferred as being less or more than the reference group. Thus, estimates with 95% CIs that did not overlap zero were considered as being statistically significant.