diff --git a/src/loading/variable.jl b/src/loading/variable.jl
index c412ad8..97d596c 100644
--- a/src/loading/variable.jl
+++ b/src/loading/variable.jl
@@ -71,30 +71,6 @@ end
 @inline _construct(cdf, name, vdr, dims::NTuple{N, Int}, ::Type{T}) where {N, T} =
     CDFVariable{T, N, typeof(vdr), typeof(cdf)}(name, vdr, cdf, dims)
 
-"""
-    read!(ds::CDFDataset, name, dest::AbstractArray{T, N}) -> dest
-
-Read the full contents of variable `name` into the preallocated `dest`.
-"""
-function Base.read!(ds::CDFDataset, name::String, dest::AbstractArray{T, N}) where {T, N}
-    vdr = find_vdr(ds, name)
-    isnothing(vdr) && throw(KeyError(name))
-    return _read_full!(dest, ds, name, vdr)
-end
-
-"""
-    read(ds::CDFDataset, name, ::Type{Array{T, N}}) -> Array{T, N}
-
-Allocating variant of [`read!`](@ref): read the full contents of variable `name` into a
-freshly allocated `Array{T, N}`.
-"""
-function Base.read(ds::CDFDataset, name::String, ::Type{Array{T, N}}) where {T, N}
-    vdr = find_vdr(ds, name)
-    isnothing(vdr) && throw(KeyError(name))
-    dims = (map(Int, record_sizes(vdr, ds, Val(N - 1)))..., Int(vdr.max_rec) + 1)
-    return _read_full!(Array{T, N}(undef, dims), ds, name, vdr)
-end
-
 function _read_full!(dest::AbstractArray{T, N}, ds, name, vdr) where {T, N}
     Base.require_one_based_indexing(dest)
     Tfile = julia_type(vdr.data_type, vdr.num_elems)
diff --git a/src/records/vdr.jl b/src/records/vdr.jl
index 4513fcf..36f8652 100644
--- a/src/records/vdr.jl
+++ b/src/records/vdr.jl
@@ -74,9 +74,6 @@ Load an r-Variable Descriptor Record from the buffer at the specified offset.
     return rVDR{FieldSizeT}(fields..., pos)
 end
 
-
-# Static-arity variants for the typed `read!` path: the caller supplies the dimension
-# count via `Val`, so tuple lengths stay inferable under `juliac --trim`.
 function record_sizes(vdr::VDR, cdf, ::Val{M}) where {M}
     vdr.num_dims == M ||
         throw(DimensionMismatch("variable has $(vdr.num_dims) dimensions, expected $M"))
diff --git a/test/cdf_trim_probe.jl b/test/cdf_trim_probe.jl
index 7bf6323..c62ee91 100644
--- a/test/cdf_trim_probe.jl
+++ b/test/cdf_trim_probe.jl
@@ -7,11 +7,5 @@ function (@main)(args::Vector{String})
          ds.majority == CDF.Row &&
          ds.compression == CDF.NoCompression &&
          length(keys(ds)) > 0
-    # Typed data reads: z-variable (VDR) 1D + 2D with row-major swap.
-    v = read!(ds, "var", Vector{Float64}(undef, 101))
-    ok &= v[1] == 1.0
-    m = read!(ds, "var2d_counter", Matrix{Float64}(undef, 10, 10))
-    ok &= m[1, 1] == 0.0 && m[2, 1] == 1.0 && m[10, 10] == 99.0
-    ok &= read(ds, "var2d_counter", Matrix{Float64}) == m
     return ok ? 0 : 1
 end
diff --git a/test/runtests.jl b/test/runtests.jl
index 6b6fcf8..0b499d1 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -71,30 +71,6 @@ end
     @test occursin("attributes: ", sprint(show, MIME("text/plain"), ds["var"]))
 end
 
-@testset "read! typed entry point" begin
-    ds = CDFDataset(data_path("a_cdf.cdf"))
-    v = read!(ds, "var", Vector{Float64}(undef, 101))
-    @test v == ds["var"][:]
-    m = read!(ds, "var2d_counter", Matrix{Float64}(undef, 10, 10))
-    @test m == reshape(0.0:99.0, 10, 10)
-    @test_throws KeyError read!(ds, "nonexistent", zeros(1))
-    @test_throws ArgumentError read!(ds, "var", zeros(Float32, 101))
-    @test_throws DimensionMismatch read!(ds, "var", zeros(100))
-    @test_throws DimensionMismatch read!(ds, "var", zeros(101, 1))
-
-    @test read(ds, "var", Vector{Float64}) == v
-    @test read(ds, "var2d_counter", Matrix{Float64}) == m
-    @test_throws KeyError read(ds, "nonexistent", Vector{Float64})
-    @test_throws ArgumentError read(ds, "var", Vector{Float32})
-    @test_throws DimensionMismatch read(ds, "var", Matrix{Float64})
-
-    rds = CDFDataset(data_path("ac_h0_mfi_20230102_v07.cdf"))
-    rvar = rds["BGSEc"]
-    b = read!(rds, "BGSEc", Matrix{eltype(rvar)}(undef, size(rvar)))
-    @test b == rvar[:, :]
-    @test read(rds, "BGSEc", Matrix{eltype(rvar)}) == b
-end
-
 @testset "Compressed cdf file (gzip)" begin
     compressed = CDFDataset(data_path("a_compressed_cdf.cdf"))
     reference = CDFDataset(data_path("a_cdf.cdf"))
diff --git a/test/trim.jl b/test/trim.jl
index 72614c0..22c8664 100644
--- a/test/trim.jl
+++ b/test/trim.jl
@@ -5,10 +5,8 @@
 # stray dynamic dispatch (e.g. a record-size type passed as a value instead of `::Type{T}`,
 # or `getproperty` dragging in the heterogeneous attribute machinery) fails the verifier.
 #
-# Data reading is covered through the typed entry point `read!(ds, name, dest)`, where
-# eltype/ndims come from `dest` instead of the file. The generic `ds[name]` path stays
-# out of scope: its element type is only known at runtime, so it cannot be statically
-# resolved.
+# The generic `ds[name]` path stays out of scope: its element type is only known at runtime,
+# so it cannot be statically resolved.
 
 using Test