larrows error inside panel.groups function

[gavril0]

I get an error with when plotting confidence intervals with larrows in a plot with a grouping variable (but it works with lsegments).

newdat <- read.table(text="
    age    Sex distance     pvar1    tvar1      thi       tlo
     8 Female 21.20909 0.4030807 6.177568 33.31712  9.101058
    10 Female 22.16818 0.3418827 6.116370 34.15627 10.180096
    12 Female 23.12727 0.3667087 6.141196 35.16402 11.090529
    14 Female 24.08636 0.4775585 6.252046 36.34037 11.832354
     8   Male 22.61563 0.2771180 6.051605 34.47677 10.754479
    10   Male 24.18438 0.2350444 6.009532 35.96306 12.405693
    12   Male 25.75312 0.2521122 6.026600 37.56526 13.940990
    14   Male 27.32187 0.3283215 6.102809 39.28338 15.360370
  ", header=TRUE)

I get the desired plot when the vertical bars are drawn with panel.segments:

xyplot(distance ~ age, group= Sex, data=newdat, ylim=c(0,50),
   panel = panel.superpose.2, col=c("red","blue"),
   panel.groups = function (x,y, lx, ux, subscripts, group.value, ...){
     x <- if(group.value=="Male") x - 0.1 else x + 0.1
     panel.xyplot(x,y,...)
     panel.segments(x0 = x,  y0 = lx[subscripts], x1 = x, y1 = ux[subscripts], ...)
#    panel.arrows(x0 = x,  y0 = lx[subscripts], x1 = x, y1 = ux[subscripts], ...)
   },
  lx=newdat$tlo, ux=newdat$thi
)

but not with panel.arrows:

xyplot(distance ~ age, group= Sex, data=newdat, ylim=c(0,50),
   panel = panel.superpose.2, col=c("red","blue"),
   panel.groups = function (x,y, lx, ux, subscripts, group.value, ...){
     x <- if(group.value=="Male") x - 0.1 else x + 0.1
     panel.xyplot(x,y,...)
#   panel.segments(x0 = x,  y0 = lx[subscripts], x1 = x, y1 = ux[subscripts], ...)
    panel.arrows(x0 = x,  y0 = lx[subscripts], x1 = x, y1 = ux[subscripts], ...)
   },
  lx=newdat$tlo, ux=newdat$thi
)

=> Errors using packet 1: invalid 'ends' or 'type' argument

I have tried to specify arrows arguments (ends, length, angle) but it does not help. Note that the function larrows works in the panel function:

xyplot(distance ~ age | Sex, data=newdat,  ylim=c(0,50), 
   panel = function (x,y, lx, ux, subscripts, ...) {
     panel.xyplot(x,y,...)
      panel.arrows(x0 = x,  y0 = lx[subscripts], x1 = x, y1 = ux[subscripts], 
      ends="both", length=0.1, angle=90, ...)
   },
  lx=newdat$tlo, ux=newdat$thi
)

Session info: R version 4.3.1, lattice_0.21-8