Insect Ecology Data Analysis Toolkit

A collection of analytical tools for insect ecology research, currently covering age-stage, two-sex life table analysis, dose-response bioassays, temperature-dependent development and insect phenology prediction. The life table module follows the age-stage, two-sex life table theory of Chi (1988) and Chi et al. (2020) . It supports fast batch processing of multi-group datasets, validates raw 'csv' data, computes cohort size, mean fecundity, age-stage survival rates, age-specific survival, age-specific fecundity, life expectancy, and derived population parameters (net reproductive rate, intrinsic and finite rates of increase, mean generation time), simultaneously generates age-stage survival curves for all groups, and exports all tabular results and plots to 'Excel' in a single run. The bioassay module estimates lethal concentrations by the traditional and the weighted (improved) linear regression methods and by probit analysis (Bliss, 1934) , with the control-mortality correction of Abbott (1925) , 95% confidence intervals and chi-square goodness-of-fit tests; the lethal proportion can be set freely (e.g., 25%, 50%, 70% or 90%), so any LC value such as the LC25, LC70 or LC90 can be computed, not only the LC50. The regression plots and tables are exported to 'Excel'. The degree-day module estimates the developmental threshold temperature and the effective accumulated temperature by the linear degree-day method (Campbell et al., 1974) , and fits the common nonlinear temperature- dependent development models following Logan et al. (1976) (Logan-6), Lactin et al. (1995) and Briere et al. (1999) , plus a 7-parameter Wang model; it selects the best model by AICc, predicts durations and accumulates field degree-days. The emergence module applies the stage-grading method to a single survey of the population stage structure, i.e. to stage-frequency data (Kiritani and Nakasuji, 1967) and Manly (1974) , and projects the beginning, peak and end of the adult emergence period (the 16%, 50% and 84% quantiles), optionally shifting the eclosion dates by the pre-oviposition period and the egg duration to forecast larval hatch. Further extensions, such as median lethal temperature/time (LT50), are planned.

insectecol

Introduction

insectecol (Insect Ecology Data Analysis Toolkit) is a collection of analytical tools for insect ecology research. It currently ships four modules:

• Age-stage, two-sex life table - main function lifeTable_analyze() for data already loaded in R, batch function lifeTable_calculate() for csv files on disk. Validates raw csv data, computes the cohort size (N), mean fecundity (F), age-stage survival rates (s_xj), age-specific survival (l_x), age-specific fecundities (F_xj, m_x), life expectancy (e_x) and the derived population parameters (net reproductive rate R0, intrinsic and finite rates of increase r and lambda, mean generation time T), draws the age-stage survival curves and exports all tabular results and plots to Excel in a single run. Fast batch processing of multi-group datasets is supported.
• Dose-response bioassay - main function lc50_analyze() for data already loaded in R, one-step batch functions lc50_export_auto() (Excel tables) and lc50_export_plot_auto() (figures) for csv files on disk. Estimates lethal concentrations by the traditional and the weighted (improved) linear regression methods and by probit analysis, with Abbott correction, 95% confidence intervals and chi-square goodness-of-fit tests. The lethal proportion can be set freely (25%, 50%, 70%, 90%, ...), so any LC value such as the LC25, LC70 or LC90 can be computed - not only the LC50. Regression plots and tables are exported to Excel.
• Degree-day / thermal constants - main function gdd_analyze() for data already loaded in R (column vectors, a data frame, or csv/xlsx file(s)). Estimates the developmental threshold temperature C and the effective accumulated temperature K by the linear degree-day law (T = C + K·V, fitted with the exact standard errors of the linear model), fits six common nonlinear temperature-dependent development models (Logan-6, Lactin 1995, Briere-1/2 1999, Wang-7) and selects the best per group by AICc (model = "auto"), with a linear-range check that warns when the rate declines at high temperatures. Publication figures use a serif font by default (Times New Roman on Windows; Chinese characters fall back to SimSun automatically) and can be exported at any physical size and resolution. Further tools: prediction (gdd_predict()), pairwise group comparison (gdd_compare()), degree-day accumulation from daily Tmin/Tmax (gdd_daily()), deriving data from a life-table csv csv/xlsx export (gdd_export()) and png export (gdd_export_plot()).
• Emergence-period projection (stage-grading method) - main function emergence_analyze() for data already loaded in R (column vectors, a data frame, or csv/xlsx file(s)). Turns one field survey of the population stage structure (e.g. a dissected-sample count of pupal grades) into the projected dates of the 16% / 50% / 84% emergence quantiles - the beginning, peak and end of the adult emergence period - by the classic Chinese stage-grading method (分龄分级推算法), and optionally projects the larval hatch dates from the pre-oviposition period and the egg duration. Quantiles outside the surveyed range are extrapolated with an explicit warning, English and Chinese column headers are auto-detected, publication figures follow the same serif-font conventions as the degree-day module, and tables and figures are exported with emergence_export() / emergence_export_plot().

Which main function should I use?

Each module has a main function for analysing data that are already loaded in R (a data frame or plain vectors) and a one-step batch function for processing csv files on disk:

The main functions assemble the data, compute everything and optionally build the plots, but never write to disk - export is handled separately by lifeTable_export(), lc50_export() and lc50_export_plot(), so the results stay fully customisable inside R.

For full control, every module can also be driven step by step (lifeTable_read() -> lifeTable_calculate_all() -> lifeTable_plot() -> lifeTable_export(), and lc50_read() -> lc50_calculate() -> lc50_plot() -> lc50_export() / lc50_export_plot()); see the function reference below.

Planned extensions include more insect ecology indicators, such as the median lethal temperature/time (LT50).

Installation

# from CRAN (once accepted)
install.packages("insectecol")

# development version from GitHub
# install.packages("devtools")
devtools::install_github("SeaGhost-0/insectecol")

Quick start: life table

library(insectecol)

# example data shipped with the package
f <- system.file("extdata", "lifetable_example.csv", package = "insectecol")
d <- read.csv(f)

# analyse straight from the columns of the loaded data frame
out <- lifeTable_analyze(
  stages = d[2:8],           # one column per immature stage
  adult_days = d$Adult,      # adult survival days
  sex = d$gender,            # "F" / "M" / "N" (died before adult)
  oviposition = d[, 11:17],  # daily oviposition of the females
  file_name = "Example"
)

out$results$N        # cohort size
out$results$R0       # net reproductive rate
out$results$lambda   # finite rate of increase

# survival analysis only: skip the reproduction-related parameters
out2 <- lifeTable_analyze(
  stages = d[2:8],
  adult_days = d$Adult,
  sex = d$gender,
  fecundity = FALSE          # no oviposition data required
)

# with the age-stage survival curve (a ggplot object)
out3 <- lifeTable_analyze(
  stages = d[2:8], adult_days = d$Adult, sex = d$gender,
  oviposition = d[, 11:17], plot = TRUE
)
print(out3$plot)

To batch-process csv files on disk instead (each csv gets its own Excel workbook with all results and the survival curve; an additional all.xlsx summarises every file):

lifeTable_calculate("path/to/lifetable_data")

Quick start: bioassay (LC)

library(insectecol)

# three parallel vectors - no csv file involved
conc   <- c(0, 1.5, 3, 6, 12, 24)
tested <- c(120, 60, 60, 60, 60, 60)
dead   <- c(7, 9, 18, 32, 48, 57)

out <- lc50_analyze(
  concentration = conc,
  tested = tested,
  dead = dead,
  name = "trial1",
  method = "all",            # traditional + improved + probit in one call
  lc = 0.5                   # LC50; any proportion works (e.g. 0.9 = LC90)
)

out$results$summary_df       # estimate, 95% CI, slope, chi-square, ...

# ... or straight from the example csv shipped with the package
f <- system.file("extdata", "lc50_example.csv", package = "insectecol")
out_csv <- lc50_analyze(lc50_read(f), method = "all")
out_csv$results$summary_df

# with the regression plot (a named list of ggplot objects)
out2 <- lc50_analyze(
  concentration = conc, tested = tested, dead = dead,
  name = "trial1", method = "probit", plot = TRUE
)
print(out2$plot$trial1)

To batch-process csv files on disk instead (one xlsx / one tiff per csv, written next to the raw data; non-default settings are appended to the file names, e.g. LB_48_LC90_probit.xlsx):

lc50_export_auto("path/to/bioassay_data", method = "probit")
lc50_export_plot_auto("path/to/bioassay_data", method = "probit")

Quick start: degree-day (thermal constants)

library(insectecol)

# example data shipped with the package
f <- system.file("extdata", "gdd_example.csv", package = "insectecol")
d <- read.csv(f)

# analyse straight from the columns of the loaded data frame
out <- gdd_analyze(
  temp     = d$temp,      # temperature (deg C)
  duration = d$duration,  # mean developmental duration (days)
  group    = d$stage      # optional grouping column (e.g. life stage)
)

out$fit$results    # C, K, SE and 95% CI per group (linear degree-day law)
summary(out$fit)   # detailed coefficient tables

# nonlinear models with AICc selection and a publication png
out2 <- gdd_analyze(temp = d$temp, duration = d$duration, group = d$stage,
                    model = "auto",          # best of six models per group
                    plot = TRUE, plot_file = "gdd.png",
                    plot_units = "cm", plot_width = 16, plot_res = 300)
out2$fit$comparison   # full model comparison table, best flag included

Without plot_file the figure is drawn on the current device and stays fully customisable via gdd_plot() (custom titles/axis labels, named per-group titles, family font). The exported png size is physical (plot_units = "in"/"cm"/"px") so plot_res only changes the sharpness and the recorded dpi, never the layout.

For batch csv processing, gdd_analyze(path = "folder") reads every csv/xlsx in a folder (combined with a source_file column), and gdd_export() writes the result tables (gdd_export_plot() the figure):

out3 <- gdd_analyze(path = "path/to/gdd_data", model = "auto")
gdd_export(out3$fit, file = "gdd_results.csv")
gdd_export_plot(out3$fit, file = "gdd.png")

Quick start: emergence period (stage-grading method)

library(insectecol)

# example data shipped with the package: one survey of the pupal
# grade structure (Tianyang overwintering generation, 40 individuals)
f <- system.file("extdata", "emergence_example.csv", package = "insectecol")
d <- read.csv(f)
#   stage        count  days     # days = days from this stage to
#   Pupal exuviae   2      0     # adult eclosion at the current
#   Pupa 7          3      2     # temperature (most developed first)
#   ...

# analyse straight from the columns of the loaded data frame
out <- emergence_analyze(
  stage = d$stage, count = d$count, days = d$days,
  survey_date = "2026-03-20"
)

out$fit$predictions   # dates of the beginning (16%), peak (50%)
                      # and end (84%) of the emergence period
predict(out$fit, c(0.25, 0.75))          # arbitrary quantiles
summary(out$fit)                         # full cumulative table

# larval hatch: eclosion + pre-oviposition period + egg duration
out2 <- emergence_analyze(data = d, survey_date = "2026-03-20",
                          pre_ovip = 3, egg_days = 10,
                          plot = TRUE, plot_file = "emergence.png")
out2$fit$predictions$hatch_date

The three quantile dates are interpolated on the cumulative development curve built from the survey; a survey that misses stages simply renormalises the shares, while a quantile below the share of the most developed stage (partly eclosed before the survey) is extrapolated backwards with an explicit warning. Tables are exported with emergence_export(fit, file = "emergence_results.csv"), the figure with emergence_export_plot(fit, file = "emergence.png").

Example data

Four example csv files ship with the package in inst/extdata/; the examples in this README and in the help pages are built on them:

system.file("extdata", "lifetable_example.csv", package = "insectecol")      # life table
system.file("extdata", "lc50_example.csv", package = "insectecol")     # bioassay
system.file("extdata", "gdd_example.csv", package = "insectecol")  # degree-day
system.file("extdata", "emergence_example.csv",
            package = "insectecol")                                # emergence

• Example.csv - life table data in the csv template: one row per individual with the ID, the days spent in each of the seven immature stages (egg, four larval instars, prepupa, pupa), the adult survival days, the sex (F/M/N) and the daily oviposition of the females.
• bioassay.csv - bioassay data: one row per concentration group with the columns Concentration (0 = control group for the Abbott correction), Tested and Dead.
• gdd_example.csv - degree-day data: one row per temperature with the columns temp (deg C), duration (mean developmental duration in days) and stage (the grouping column). inst/extdata/gdd_batch/ additionally ships one file per temperature for the batch mode.
• emergence_example.csv - emergence-period survey data (the Tianyang overwintering-generation case): one row per stage with the columns stage, count (individuals in that stage) and days (average days from that stage to adult eclosion).

The file layouts are described in detail under Data formats.

Data formats

Life table csv

One row per individual. If the sex column is at position n:

The first line must contain the stage names as headers. The sex column is located automatically, and the file encoding is detected automatically (UTF-8 and GBK are both supported).

Bioassay csv

One row per concentration group (replicates = repeated concentration values):

Headers are matched loosely, so a header like Concentration (mg/L) is recognised as well. UTF-8 (with BOM) and GBK encodings are supported.

Degree-day csv

One row per temperature (long format):

The temperature and duration columns are auto-detected (English and Chinese headers are recognised); ambiguous files accept explicit temp_col / duration_col. UTF-8 (with BOM) and GBK encodings are supported; the csv delimiter (, ; tab) is auto-detected, and xlsx files are read via readxl.

Emergence csv

One row per stage, ordered most-developed-first (the rows are sorted by days internally anyway):

The columns are auto-detected (English and Chinese headers are recognised); ambiguous files accept explicit stage_col / count_col / percent_col / days_col. Exactly one of count / percent is used (count wins with a message when both are present). UTF-8 (with BOM) and GBK encodings are supported; the csv delimiter is auto-detected, and xlsx files are read via readxl.

Function reference

Life table module

Bioassay module

Degree-day module

Emergence module

Updates

1.1.2 (CRAN submission round)

Figures and fonts

• In a label that mixes the two scripts, every character now keeps its own font: digits, symbols and Latin words stay in Times New Roman and only the Chinese characters are set in the system CJK font (SimSun on Windows, Songti on macOS). A label used to be set in one font as a whole, so "Concentration (mg/L)" written with a Chinese unit came out entirely in the CJK font, Latin letters included.
• Text sizes are true typographic points on every device and at every dpi, so a label has the same physical size in a 300 dpi png and in a pdf (it used to grow with the dpi).
• Chinese labels are written to vector devices (pdf, svg) as well: the TrueType font is embedded, and the family is registered with the classic font databases, which removes the "invalid font type" and "font family not found" failures when a figure is printed to the default device.
• The LC label of lc50_plot() falls back to plain text when the unit contains Chinese, so that the unit is set in the CJK font like every other label. The percentage is then written LC50 instead of with a subscript, because plotmath can only draw a whole expression with one font.
• plot = TRUE in lifeTable_analyze(), lc50_analyze(), emergence_analyze() and gdd_analyze() now always writes the figure: with no plot_file it goes to the working directory under a default name (<file_name>_plot.png, LC50_<name>.png, emergence_plot.png or gdd_plot.png), and a plot_file without an extension is treated as a folder - created when missing - with the figure written inside it. Previously the figure was only returned and ggsave() stopped on a folder path.
• The emergence and degree-day figure devices follow the plot_file extension (png/tiff/jpeg written by ragg when available).

Examples

• \dontrun replaced by \donttest throughout: every example can be run by the user, and only the parts that write a workbook - which take well over five seconds - are skipped by R CMD check.

Internal changes

• Column assignment in the age-stage life table now follows the position of each duration in the data row instead of assuming that the last entry is the adult stage: individuals that died before the adult stage and were sexed F/M (trailing blanks) and rows with a skipped stage (blank cell, data in later columns) are no longer misplaced into the Female/Male columns of s_xj and e_xj.
• Figures exported by lifeTable_analyze(plot = TRUE) are now written through the same device machinery as the other functions: bitmaps go through ragg (per-glyph font fallback) and vector formats through showtext, so a plot_file pointing at a pdf or svg works as well.
• Device font registration no longer calls the Windows-only windowsFonts(), and the bundled-font fallback registers its metrics with the vector devices too - the package builds and checks cleanly on Linux and macOS.
• systemfonts added to Imports (the package ragg draws and measures text with); it supplies the per-character widths used to split a mixed label into runs.
• Bumped the version to 1.1.2.

1.1.1 (CRAN submission round)

Version 1.1.1 adds the emergence-period module, standardises the function naming across all modules (<module>_<verb> with the same verbs everywhere, see below) and removes the no-longer-needed gdd_from_lifetable() bridge.

Function naming unified across modules

• Every pipeline function now follows the <module>_<verb> scheme of the newer modules, and every module uses the same word for the same action - *_analyze() (main entry), *_read() (file intake), *_plot() (figures), *_export() (tables to csv/xlsx), *_export_plot() (figures to png/tiff). Renamed: read_life_table() -> lifeTable_read(), build_life_table() -> lifeTable_build(), plot_sxj() -> lifeTable_plot(), check_life_table() -> lifeTable_check(), save_results() -> lifeTable_export(), read_lc50() -> lc50_read(), plot_lc50() -> lc50_plot(), save_lc50() -> lc50_export(), save_lc50_plot() -> lc50_export_plot(), save_lc50_auto() -> lc50_export_auto() and save_lc50_plot_auto() -> lc50_export_plot_auto(). The old names are removed (the previous release had essentially no users); the low-level indicator helpers (calc_N(), calc_R0(), ...) keep their short names.
• New gdd_export_plot() and emergence_export_plot(): export the figure of an existing fit as png at any time (the standalone counterpart of plot_file = in the two *_analyze() functions, which now reuse them internally).
• plot_file png export added to all four main functions: the life-table (lifeTable_analyze()) and bioassay (lc50_analyze()) main functions now also accept plot_file (with plot_width/plot_height/plot_units/plot_res), so one call goes from raw data to the finished figure file in every module.
• Removed gdd_from_lifetable(): the raw-life-table bridge is not needed any more.
• Emergence figure reworked for manuscript use: text sizes scale with the new cex argument (default 2 - at half the text width of a manuscript the labels read at about the body-text size), lines are thicker (lwd), the quantile legend moved to a framed box on the right-hand side (one block per quantile, with a true arrow glyph instead of "->"), the survey annotation sits above the x axis, and the axis-title spacing no longer clips.

New module: emergence-period projection (stage-grading method)

• New main function emergence_analyze(): turns one field survey of the population stage structure (e.g. a dissected-sample count of pupal grades) into the projected dates of the 16% / 50% / 84% emergence quantiles (the beginning, peak and end of the adult emergence period, i.e. the mean +/- 1 SD of a normal emergence curve) by the classic Chinese stage-grading method. Accepts column vectors, a data frame, or csv/xlsx file(s)/folder; nothing is written to disk unless plot_file is supplied.
• Larval hatch projection: pre_ovip (pre-oviposition period) and egg_days (egg duration) shift the eclosion dates to the hatch dates, e.g. for forecasting the hatch of larvae from a pupal-grade survey.
• Quantiles at or below the cumulative share of the most developed stage (partly eclosed before the survey) are extrapolated backwards from the first segment with an explicit warning; the survey shares are renormalised when stages are missing. Zero-count stages are dropped, rows are sorted by days to eclosion automatically, and stages sharing one days value are flagged for checking.
• Column auto-detection with English and Chinese aliases (stage/虫态, count/数量, days/历期, ...), delimiter and encoding handling as in the degree-day module.
• predict(fit, p) interpolates arbitrary quantiles; the fit$interpolate closure supports further programming.
• Publication figures with the same serif-font conventions and physical-size png export as the degree-day module; emergence_export_plot() writes the figure, emergence_export() writes the prediction and stage tables to csv/xlsx.
• Example data: inst/extdata/emergence_example.csv (Tianyang overwintering-generation survey, 40 individuals, 10 stages).

New module: degree-day / thermal constants

• New main function gdd_analyze(): one call from column vectors (temp = d$temp, duration = d$days, group = d$stage), a data frame, or csv/xlsx file(s)/folder - with optional data validation (gdd_check()), model fitting and png export. Nothing is written to disk unless plot_file is supplied.
• Linear degree-day law fitted as T = C + K·V, so the threshold temperature C and the effective accumulated temperature K carry the exact standard errors and confidence intervals of the linear model (verified to match SPSS to every digit).
• Six nonlinear temperature-dependent development models (Logan-6, Lactin 1995, Briere-1/2 1999, Wang-7) with a robust restart strategy and model = "auto" selecting the best model per group by AICc.
• Linear-range check: warns when the developmental rate declines at high temperatures (strong warning for model = "linear", mild for "auto"/nonlinear).
• Publication figures: serif font by default (family = "serif", which is Times New Roman on Windows) with per-glyph fallback (Chinese characters render in SimSun on Chinese Windows, no showtext required); custom titles/subtitles/axis labels including named per-group titles; physical-size png export (plot_units = "in"/"cm"/"px", plot_res dpi) where the resolution changes only the sharpness, never the layout.
• Further tools: gdd_read() (csv/xlsx file or folder, delimiter and column auto-detection, English/Chinese headers), gdd_compare() (pairwise group comparison), gdd_predict(), gdd_daily() (avg and triangle methods), gdd_export() (csv/xlsx) and gdd_export_plot() (png figure).
• Example data: inst/extdata/gdd_example.csv and the per-temperature files in inst/extdata/gdd_batch/.

New: bootstrap for the life table

• New lifeTable_bootstrap(): nonparametric individual-level bootstrap of the life table (default B = 100000) with percentile confidence intervals for the population parameters, using a fully vectorised Euler-Lotka solver. The result is attached as results$boot when lifeTable_analyze(bootstrap = TRUE) is called (arguments B, seed), following the TWOSEX-MSChart technique.
• New lifeTable_boot_test(): paired two-cohort bootstrap comparison.

Improved

• LC50 regression plots: when a replicate error bar is wide enough to reach into the LC reference label, the label now shifts vertically by the smallest amount that restores a clearance of about one line height from the bar end (preferring its own side of the LC crossing) instead of staying at its fixed height and colliding; when bars crowd it from above and below, it slides sideways to the nearest free spot. The x-axis value of the dashed line always sits to the right of the line (it moves left only when it would run off the panel edge); a label block that would cover it is raised clear instead.
• Plots saved with non-default display options now get a suffix in the file name (_linear, _noband, _nobar, _nolcCI, _nochi), so different settings saved to one folder never overwrite each other.

Internal changes

• Regenerated the roxygen documentation and NAMESPACE for all new and renamed functions and S3 methods (print/summary/plot/predict methods for the gdd and emergence objects, print for the bootstrap object).
• Bumped the version to 1.1.1.

1.0.1 (CRAN submission round)

New features

• New main functions lifeTable_analyze() and build_life_table() for the life table module: analyse data already loaded in R - pass the stage columns, adult days, sex and oviposition columns of any data frame; no package-conform csv file required. Nothing is written to disk.
• New main function lc50_analyze() for the bioassay module: accepts a data frame, a named list of data frames or three parallel vectors (concentration, tested, dead) and computes the LC values with the selected method(s).
• New one-step batch exporters save_lc50_auto() and save_lc50_plot_auto(): read every csv in a folder, compute the LC values and write the xlsx/tiff results next to the raw data, with self-documenting file names (e.g. LB_48_LC90_probit.xlsx).
• lc50_calculate() now accepts several methods (or method = "all") in a single call and reports the status of every method for every file.
• Redesigned LC regression plots: replicate rows pooled with Wilson score intervals, pointwise confidence band of the fitted curve, dashed LC reference lines with automatic tick and label placement, and a shape argument switching between the log10 (sigmoid) and the linear concentration axis. The LC reference label now also shows the 95% confidence interval of the estimate on a second line, e.g. LC50 = 1.23 mg/L over (0.98-1.55), and optionally the chi-square goodness-of-fit result on a third line (lc_ci = FALSE / lc_p = FALSE omit the lines; lc_lab_gap (left of the reference line) / lc_lab_gap_right (right of it) / lc_lab_dy fine-tune the label position, lc_lab_lh its line spacing, given in multiples of the font size). Every line of the label is drawn on its own so that all of them share one vertical axis (grid would otherwise justify each line of a multi-line string by its own width), and the label is kept inside the panel. The block is placed in the diagonal quadrant around the crossing that the rising fitted curve never enters - above it when the label sits left of the vertical reference line, below it when it sits right - anchored by the edge facing the crossing, so adding or dropping a line (or changing lc_lab_lh) grows the block away from the crossing instead of onto the dashed line or the curve. The LC estimate is additionally marked by a circle where it lies on the fitted curve, i.e. where the two dashed reference lines meet.
• Customisable titles, axis titles and legend labels for the age-stage survival curves (plot_sxj() and lifeTable_analyze()).

CRAN fixes

• Quoted the software names in DESCRIPTION ('csv', 'Excel').
• Unwrapped the example code.
• Replaced cat() with message() for all progress output.
• showtext is now enabled at export time and also works for the pdf device (vector devices are converted at 72 pt/in); the previous session settings are restored afterwards, so exports look identical in every R session.
• Bumped the version to 1.0.1.

Internal changes

• lc50_traditional(), lc50_improved() and lc50_probit() are no longer exported; select them via the method argument of lc50_calculate() / lc50_analyze().
• Font fallback chain for publication-quality figures: system Times New Roman -> bundled Liberation Serif -> "sans" (see the font note below).
• The oviposition consistency check is skipped gracefully when the data end at the sex column (no oviposition columns).
• Declared R (>= 3.5) and LazyData in DESCRIPTION.
• Added a GitHub Actions R-CMD-check workflow (macOS / Windows / Linux: release, devel and oldrel).
• Updated the CITATION file.
• Added example data for the planned thermal-constants module (inst/extdata/gdd_example.csv).
• Normalised the source indentation to two spaces and rewrote the README (main-function guide, example data section, quick starts on the shipped csv files).

1.0.0

• Initial release: the age-stage, two-sex life table module and the dose-response bioassay (LC) module.

Note on bundled fonts

This package bundles the Liberation Serif font (SIL Open Font License 1.1) in inst/fonts/ for publication-quality figures. The full license text is shipped as inst/fonts/OFL.txt. All other components of the package are licensed under MIT.

License

MIT (see LICENSE). The bundled Liberation Serif font is licensed under the SIL Open Font License 1.1.

Citation

citation("insectecol")

If you use the life table module in a publication, please also cite the method papers behind the age-stage, two-sex theory (Chi & Liu 1985; Chi 1988 - see the references of lifeTable_read()), and for probit analysis Finney (1971) together with Abbott (1925) for the correction of natural mortality.