diff --git a/src/fasta/readrecord.jl b/src/fasta/readrecord.jl
index 7f84c6c..de2eb63 100644
--- a/src/fasta/readrecord.jl
+++ b/src/fasta/readrecord.jl
@@ -12,11 +12,11 @@ machine = (function ()
     identifier.actions[:enter] = [:pos]
     identifier.actions[:exit]  = [:identifier]
     
-    description = re.cat(re.any() \ hspace, re"[^\r\n]*")
+    description = re.cat(re.any() \ re.space(), re"[^\n\r]*")
     description.actions[:enter] = [:pos]
     description.actions[:exit]  = [:description]
-    
-    header = re.cat('>', identifier, re.opt(re.cat(re.rep1(hspace), description)))
+
+    header = re">" * identifier * re.opt(re.rep1(hspace) * re.opt(description))
     header.actions[:enter] = [:mark]
     header.actions[:exit]  = [:header]
     
@@ -117,4 +117,4 @@ Automa.Stream.generate_reader(
     initcode = initcode,
     loopcode = loopcode,
     returncode = returncode
-) |> eval
\ No newline at end of file
+) |> eval
diff --git a/src/fastq/readrecord.jl b/src/fastq/readrecord.jl
index 8cbd53f..5471c6a 100644
--- a/src/fastq/readrecord.jl
+++ b/src/fastq/readrecord.jl
@@ -10,11 +10,11 @@ machine = (function ()
         identifier.actions[:enter] = [:pos]
         identifier.actions[:exit]  = [:header1_identifier]
         
-        description = re.cat(re.any() \ hspace, re"[^\r\n]*")
+        description = re.cat(re.any() \ re.space(), re"[^\r\n]*")
         description.actions[:enter] = [:pos]
         description.actions[:exit]  = [:header1_description]
         
-        re.cat('@', identifier, re.opt(re.cat(re.rep1(hspace), description)))
+        re.cat('@', identifier, re.opt(re.cat(re.rep1(hspace), re.opt(description))))
     end
     
     sequence = re"[A-z]*"
diff --git a/test/runtests.jl b/test/runtests.jl
index 84b9c09..6836077 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -76,6 +76,11 @@ import BioSequences:
         @test FASTA.header(record) == "CYS1_DICDI fragment"
         @test FASTA.sequence(record) == aa"SCWSFSTTGNVEGQHFISQNKLVSLSEQNLVDCDHECMEYEGE"
         @test FASTA.sequence(record, 10:15) == aa"NVEGQH"
+
+        # PR 37
+        s = ">A \nTAG\n"
+        rec = first(iterate(FASTA.Reader(IOBuffer(s))))
+        @test isempty(rec.description)
     end
 
     output = IOBuffer()
@@ -291,6 +296,11 @@ end
     +2 high quality
     IJN
     """
+
+    # PR 37
+    s = "@A \nTAG\n+\nJJJ\n"
+    rec = first(iterate(FASTQ.Reader(IOBuffer(s))))
+    @test isempty(rec.description)
     
     # Test issue 25
     lines = ["@A", "A", "+", "F", "@B", "CA", "+", "CF"]