diff --git a/DESCRIPTION b/DESCRIPTION
index 9e93908..0c2eae2 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -1,7 +1,7 @@
 Package: atime
 Type: Package
 Title: Asymptotic Timing
-Version: 2026.4.2
+Version: 2026.4.5
 Depends: R (>= 3.5.0)
 Authors@R: c(
     person("Toby", "Hocking",
diff --git a/NEWS b/NEWS
index 98d6ec7..a9610bf 100644
--- a/NEWS
+++ b/NEWS
@@ -1,3 +1,7 @@
+Changes in version 2026.4.5
+
+- atime() gains check arg (default FALSE), passed to bench::mark().
+
 Changes in version 2026.4.2
 
 - atime_pkg() now sorts tests in facet plots to emphasize speedups, in addition to slowdowns. New pred.Nx used instead of N.factor (issue109 data and fix), as well as alternative="two.sided" instead of "greater" in t.test().
diff --git a/R/atime.R b/R/atime.R
index d7743fb..e62c662 100644
--- a/R/atime.R
+++ b/R/atime.R
@@ -21,8 +21,8 @@ get_result_rows <- function(result.list){
   }
 }
 
-run_bench_mark <- function(times, sub.elist, N.env, result){
-  m.list <- list(quote(bench::mark), iterations=times, check=FALSE)
+run_bench_mark <- function(times, sub.elist, N.env, result, check){
+  m.list <- list(quote(bench::mark), iterations=times, check=check)
   N.env$result.list <- list()
   for(expr.name in names(sub.elist)){
     expr <- sub.elist[[expr.name]]
@@ -36,9 +36,9 @@ run_bench_mark <- function(times, sub.elist, N.env, result){
   }
   m.call <- as.call(m.list)
   N.df <- suppressWarnings(eval(m.call, N.env))
-  if(result$keep){
-    N.df$result <- N.env$result.list
-  }
+  N.df$result <- if(result$keep){
+    N.env$result.list
+  }else list(NULL)
   N.df
 }
 
@@ -74,7 +74,7 @@ check_atime_inputs <- function(N, result, elist){
   list(keep=keep, fun=fun)
 }
 
-atime <- function(N=default_N(), setup, expr.list=NULL, times=10, seconds.limit=0.01, verbose=FALSE, result=FALSE, N.env.parent=NULL, ...){
+atime <- function(N=default_N(), setup, expr.list=NULL, times=10, seconds.limit=0.01, verbose=FALSE, result=FALSE, N.env.parent=NULL, check=FALSE, ...){
   kilobytes <- mem_alloc <- . <- sizes <- expr.name <- NULL
   ## above for CRAN NOTE.
   formal.names <- names(formals())
@@ -96,7 +96,7 @@ atime <- function(N=default_N(), setup, expr.list=NULL, times=10, seconds.limit=
     N.env <- new.env(parent=N.env.parent)
     N.env$N <- N.value
     eval(mc.args$setup, N.env)
-    N.df <- run_bench_mark(times, elist[not.done.yet], N.env, result)
+    N.df <- run_bench_mark(times, elist[not.done.yet], N.env, result, check)
     result.row.list <- get_result_rows(N.env$result.list)
     N.stats <- data.table(
       N=N.value, expr.name=not.done.yet, N.df
diff --git a/man/atime.Rd b/man/atime.Rd
index 92b0615..a94ff20 100644
--- a/man/atime.Rd
+++ b/man/atime.Rd
@@ -6,7 +6,7 @@
 
 \usage{atime(
  N, setup, expr.list=NULL, times=10, seconds.limit=0.01, verbose=FALSE,
- result=FALSE, N.env.parent=NULL, ...)}
+ result=FALSE, N.env.parent=NULL, check=FALSE, ...)}
 
 \arguments{
   \item{N}{numeric vector of at least two unique data sizes, default is \code{2^seq(2,20)}.}
@@ -25,7 +25,9 @@
     and seconds).
   }
   \item{N.env.parent}{environment to use as parent of environment
-  created for each data size N, or NULL to use default parent env.}
+    created for each data size N, or NULL to use default parent env.}
+  \item{check}{logical, error if results are not all equal? (default
+  FALSE) Passed to \code{\link[bench]{mark}}.}
   \item{\dots}{named expressions to time.}
 }
 
diff --git a/tests/testthat/test-CRAN.R b/tests/testthat/test-CRAN.R
index f55a9c0..43673e9 100644
--- a/tests/testthat/test-CRAN.R
+++ b/tests/testthat/test-CRAN.R
@@ -117,6 +117,17 @@ test_that("more units defined in 1 row result", {
   expect_identical(names(unit.tab), c("match.len","seconds"))
 })
 
+test_that("result=FALSE and check=TRUE", {
+  expr.list <- atime::atime_grid(seq=seq(1,N), colon=1:N)
+  alist <- atime::atime(expr.list=expr.list, foo=1)
+  alist$measurements[, expect_identical(result, replicate(.N, NULL))]
+  expect_error({
+    atime::atime(expr.list=expr.list, foo=1, check=TRUE)
+  }, "`seq` does not equal `foo`", fixed=TRUE)
+  alist <- atime::atime(expr.list=expr.list, check=TRUE)
+  alist$measurements[, expect_identical(result, replicate(.N, NULL))]
+})
+
 test_that("result returned when some are NULL and others not", {
   atime.list <- atime::atime(
     N=10^seq(-3, 0),