Index: ../src/mitk/Core/Code/DataManagement/mitkImage.h
===================================================================
--- ../src/mitk/Core/Code/DataManagement/mitkImage.h	(revision 26484)
+++ ../src/mitk/Core/Code/DataManagement/mitkImage.h	(working copy)
@@ -375,10 +375,29 @@
       }
       if(spacing[j] < mitk::eps)
       {
+        if ( (j==2) && (m_Dimensions[2] == 1) )
+        {
+          // Negative spacings can occur when reading single DICOM slices with ITK via GDCMIO
+          // In these cases spacing is not determind by ITK correctly (because it distinguishes correctly
+          // between slice thickness and inter slice distance -- slice distance is meaningless for
+          // single slices).
+          // I experienced that ITK produced something meaningful nonetheless because is is 
+          // evaluating the tag "(0018,0088) Spacing between slices" as a fallback. This tag is not
+          // reliable (http://www.itk.org/pipermail/insight-users/2005-September/014711.html)
+          // but gives at least a hint.
+          // In real world cases I experienced that this tag contained the correct inter slice distance
+          // with a negative sign, so we just invert such negative spacings.
+          MITK_DEBUG << "Illegal value of itk::Image::GetSpacing()[" << j <<"]=" << spacing[j] << ". Using inverted value " << -spacing[j];
+          spacing[j] = -spacing[j];
+        }
+        else
+        {
+          // in other cases (multiple slices or not third dimension), negative values are not expected
         MITK_ERROR << "Illegal value of itk::Image::GetSpacing()[" << j <<"]=" << spacing[j] << ". Using 1.0 instead.";
-        spacing[j] = 1.0;
+          spacing[j] = -spacing[j];
       }
     }
+    }
     if(itkdirectionOk == false)
     {
       MITK_ERROR << "Illegal matrix returned by itk::Image::GetDirection():" << itkdirection << " Using identity instead.";