diff --git a/src/java/org/apache/poi/hssf/usermodel/HSSFSheet.java b/src/java/org/apache/poi/hssf/usermodel/HSSFSheet.java
index 5855a4937e..6fc113c085 100644
--- a/src/java/org/apache/poi/hssf/usermodel/HSSFSheet.java
+++ b/src/java/org/apache/poi/hssf/usermodel/HSSFSheet.java
@@ -1615,7 +1615,13 @@ public class HSSFSheet
         for (Iterator it = rowIterator(); it.hasNext();) {
             HSSFRow row = (HSSFRow) it.next();
             HSSFCell cell = row.getCell(column);
-            if (cell == null) continue;
+
+            boolean isCellInMergedRegion = false;
+            for (int i = 0 ; i < getNumMergedRegions() && ! isCellInMergedRegion; i++) {
+                isCellInMergedRegion = getMergedRegionAt(i).contains(row.getRowNum(), column);
+            }
+
+            if (cell == null | isCellInMergedRegion) continue;
 
             HSSFCellStyle style = cell.getCellStyle();
             HSSFFont font = wb.getFontAt(style.getFontIndex());
diff --git a/src/testcases/org/apache/poi/hssf/data/43902.xls b/src/testcases/org/apache/poi/hssf/data/43902.xls
new file mode 100644
index 0000000000..a67720d6c4
Binary files /dev/null and b/src/testcases/org/apache/poi/hssf/data/43902.xls differ
diff --git a/src/testcases/org/apache/poi/hssf/usermodel/TestHSSFSheet.java b/src/testcases/org/apache/poi/hssf/usermodel/TestHSSFSheet.java
index f6f506c57a..9a6f9dca00 100644
--- a/src/testcases/org/apache/poi/hssf/usermodel/TestHSSFSheet.java
+++ b/src/testcases/org/apache/poi/hssf/usermodel/TestHSSFSheet.java
@@ -32,6 +32,7 @@ import org.apache.poi.hssf.record.VCenterRecord;
 import org.apache.poi.hssf.record.WSBoolRecord;
 import org.apache.poi.hssf.record.WindowTwoRecord;
 import org.apache.poi.hssf.util.Region;
+import org.apache.poi.poifs.filesystem.POIFSFileSystem;
 import org.apache.poi.util.TempFile;
 
 /**
@@ -614,6 +615,44 @@ public class TestHSSFSheet
         workbook = new HSSFWorkbook(new ByteArrayInputStream(out.toByteArray()));
         assertTrue("No Exceptions while reading file", true);
 
+    }
+    
+    public void testAutoSizeColumn() throws Exception {
+		String filename = System.getProperty("HSSF.testdata.path");
+		filename = filename + "/43902.xls";
+		String sheetName = "my sheet";
+		FileInputStream is = new FileInputStream(filename);
+		POIFSFileSystem fs = new POIFSFileSystem(is);
+		HSSFWorkbook wb = new HSSFWorkbook(fs);
+		HSSFSheet sheet = wb.getSheet(sheetName);
+		
+		// autoSize the first column and check its size before the merged region (1,0,1,1) is set:
+		// it has to be based on the 2nd row width
+		sheet.autoSizeColumn((short)0);
+		assertEquals("Column autosized with only one row: wrong width", (short)7169, sheet.getColumnWidth((short)0));
+		
+		//create a region over the 2nd row and auto size the first column
+		sheet.addMergedRegion(new Region(1,(short)0,1,(short)1));
+		sheet.autoSizeColumn((short)0);
+		ByteArrayOutputStream out = new ByteArrayOutputStream();
+		wb.write(out);
+		out.close();
+		
+		// check that the autoSized column width has ignored the 2nd row 
+		// because it is included in a merged region (Excel like behavior)
+		HSSFWorkbook wb2 = new HSSFWorkbook(new ByteArrayInputStream(out.toByteArray()));
+		HSSFSheet sheet2 = wb2.getSheet(sheetName);
+		assertEquals((short)3024, sheet2.getColumnWidth((short)0));
+		
+		// remove the 2nd row merged region and check that the 2nd row value is used to the autoSizeColumn width
+		sheet2.removeMergedRegion(1);
+		sheet2.autoSizeColumn((short)0);
+		out = new ByteArrayOutputStream();
+		wb2.write(out);
+		out.close();
+		HSSFWorkbook wb3 = new HSSFWorkbook(new ByteArrayInputStream(out.toByteArray()));
+		HSSFSheet sheet3 = wb3.getSheet(sheetName);
+		assertEquals((short)7169, sheet3.getColumnWidth((short)0));
     }
 
 	public static void main(java.lang.String[] args) {