diff --git a/README.md b/README.md
index 40dd079..548724c 100644
--- a/README.md
+++ b/README.md
@@ -91,6 +91,10 @@ settings:
 		bwa aln ref.fa read1.fq > read1.sai; bwa aln ref.fa read2.fq > read2.sai
 		bwa samse ref.fa reads.sai reads.fq > aln-pe.sam
 
+####Why does a read appear multiple times in the output SAM?
+
+BWA-SW and BWA-MEM perform local alignments.
+
 ####How to map sequences to GRCh38 with ALT contigs?
 
 BWA-backtrack and BWA-MEM partially support mapping to a reference containing
diff --git a/bwa.1 b/bwa.1
index 1a70a52..779835c 100644
--- a/bwa.1
+++ b/bwa.1
@@ -191,7 +191,7 @@ yields fewer seeds, which leads to faster alignment speed but lower accuracy. [1
 .BI -c \ INT
 Discard a MEM if it has more than
 .I INT
-occurence in the genome. This is an insensitive parameter. [10000]
+occurence in the genome. This is an insensitive parameter. [500]
 .TP
 .B -P
 In the paired-end mode, perform SW to rescue missing hits only but do not try to find
diff --git a/fastmap.c b/fastmap.c
index 3b782bf..c8b2eb0 100644
--- a/fastmap.c
+++ b/fastmap.c
@@ -156,7 +156,7 @@ int main_mem(int argc, char *argv[])
 		fprintf(stderr, "\n");
 		fprintf(stderr, "       -v INT        verbose level: 1=error, 2=warning, 3=message, 4+=debugging [%d]\n", bwa_verbose);
 		fprintf(stderr, "       -T INT        minimum score to output [%d]\n", opt->T);
-		fprintf(stderr, "       -h INT        if #hits < INT, output all in the XA tag [%d]\n", opt->max_hits);
+		fprintf(stderr, "       -h INT        if there are <INT hits with score >80%% of the max score, output all in XA [%d]\n", opt->max_hits);
 		fprintf(stderr, "       -a            output all alignments for SE or unpaired PE\n");
 		fprintf(stderr, "       -C            append FASTA/FASTQ comment to SAM output\n");
 		fprintf(stderr, "       -Y            use soft clipping for supplementary alignments\n");