diff --git a/src/cached.jl b/src/cached.jl
index 4c20bb5..c4f5b08 100644
--- a/src/cached.jl
+++ b/src/cached.jl
@@ -41,8 +41,8 @@ function _readblock_cached!(A::CachedDiskArray{T,N}, data, I...) where {T,N}
         if haskey(A.cache, c)
             A.cache[c]
         else
-            chunk_data = Array{T,N}(undef, length.(I)...)
-            A.cache[c] = readblock!(parent(A), chunk_data, I...)
+            chunk_data = Array{T,N}(undef, length.(c)...)
+            A.cache[c] = readblock!(parent(A), chunk_data, c...)
         end
     end
     out = ConcatDiskArray(chunk_arrays)
diff --git a/test/runtests.jl b/test/runtests.jl
index 644e577..b974dad 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -709,22 +709,24 @@ end
 end
 
 @testset "Cached arrays" begin
-    A = (1:300) * (1:1200)'
-    ch = ChunkedDiskArray((1:3000) * (1:1200)', (128, 128))
+    M = (1:300) * (1:1200)'
+    A = cat(M, M, M, M; dims=3)
+    ch = ChunkedDiskArray(A, (128, 128, 2))
     ca = DiskArrays.CachedDiskArray(ch; maxsize=5)
     # Read the original
     @test sum(ca) == sum(ca)
-    # Read from the cache
-    @test ca[:, :] == ch
     length(ca.cache)
 
     ca = DiskArrays.cache(ch; maxsize=5)
     @test sum(ca) == sum(ca)
 
-    @test ca[:, :] == ch[:, :]
-    @test ca[:, 1] == ch[:, 1]
-    @test ca[:, 2] == ch[:, 2]
-    @test ca[:, 3] == ch[:, 3]
-    @test ca[:, 200] == ch[:, 200]
-    @test ca[200, :] == ch[200, :]
+    @test ca[:, :, 1] == A[:, :, 1]
+    @test ca[:, :, 2] == A[:, :, 2]
+    @test ca[:, :, 2] == A[:, :, 3]
+    @test ca[:, :, 2] == A[:, :, 4]
+    @test ca[:, 1, 1] == ch[:, 1, 1]
+    @test ca[:, 2, 1] == ch[:, 2, 1]
+    @test ca[:, 3, 1] == ch[:, 3, 1]
+    @test ca[:, 200, 1] == ch[:, 200, 1]
+    @test ca[200, :, 1] == ch[200, :, 1]
 end