diff --git a/docs/ReleaseNotes.html b/docs/ReleaseNotes.html
index 5e1192c..d2e8cfa 100644
--- a/docs/ReleaseNotes.html
+++ b/docs/ReleaseNotes.html
@@ -52,6 +52,14 @@
     <td width="87%">Removed Variant serializable interface because it is not 
     	actually serializable on 2000/xp(pre1)</td>
   </tr> 
+  <tr>
+    <td width="13%">1224219</td>
+    <td width="87%">Memory leak in SafeArray.GetString()(pre3)</td>
+  </tr> 
+  <tr>
+    <td width="13%">1224219</td>
+    <td width="87%">Change from UTF to UNICODE SafeArray.fromStringArray()(pre3)</td>
+  </tr> 
   <tr>
     <td width="13%">&nbsp;</td>
     <td width="87%">&nbsp;</td>
diff --git a/jni/Dispatch.cpp b/jni/Dispatch.cpp
index bceee72..40b9766 100644
--- a/jni/Dispatch.cpp
+++ b/jni/Dispatch.cpp
@@ -55,6 +55,7 @@ JNIEXPORT jobject JNICALL Java_com_jacob_com_Dispatch_QueryInterface
   // get the current IDispatch
   IDispatch *pIDispatch = extractDispatch(env, _this);
   if (!pIDispatch) return NULL;
+  // if we used env->GetStringChars() would that let us drop the conversion?
   const char *siid = env->GetStringUTFChars(_iid, NULL);
   USES_CONVERSION;
   LPOLESTR bsIID = A2W(siid);
@@ -95,6 +96,7 @@ JNIEXPORT void JNICALL Java_com_jacob_com_Dispatch_createInstanceNative
   jclass clazz = env->GetObjectClass(_this);
   jfieldID jf = env->GetFieldID( clazz, DISP_FLD, "I");
 
+  // if we used env->GetStringChars() would that let us drop the conversion?
   const char *progid = env->GetStringUTFChars(_progid, NULL);
   CLSID clsid;
   HRESULT hr;
@@ -161,6 +163,7 @@ JNIEXPORT void JNICALL Java_com_jacob_com_Dispatch_getActiveInstanceNative
   jclass clazz = env->GetObjectClass(_this);
   jfieldID jf = env->GetFieldID( clazz, DISP_FLD, "I");
 
+  // if we used env->GetStringChars() would that let us drop the conversion?
   const char *progid = env->GetStringUTFChars(_progid, NULL);
   CLSID clsid;
   HRESULT hr;
@@ -203,6 +206,7 @@ JNIEXPORT void JNICALL Java_com_jacob_com_Dispatch_coCreateInstanceNative
   jclass clazz = env->GetObjectClass(_this);
   jfieldID jf = env->GetFieldID( clazz, DISP_FLD, "I");
 
+  // if we used env->GetStringChars() would that let us drop the conversion?
   const char *progid = env->GetStringUTFChars(_progid, NULL);
   CLSID clsid;
   HRESULT hr;
@@ -275,6 +279,7 @@ JNIEXPORT jintArray JNICALL Java_com_jacob_com_Dispatch_getIDsOfNames
   {
     USES_CONVERSION;
     jstring s = (jstring)env->GetObjectArrayElement(names, i);
+	// if we used env->GetStringChars() would that let us drop the conversion?
     const char *nm = env->GetStringUTFChars(s, NULL);
     LPOLESTR nmos = A2W(nm);
     env->ReleaseStringUTFChars(s, nm);
diff --git a/jni/DispatchEvents.cpp b/jni/DispatchEvents.cpp
index 54ad882..0605b02 100644
--- a/jni/DispatchEvents.cpp
+++ b/jni/DispatchEvents.cpp
@@ -81,16 +81,21 @@ JNIEXPORT void JNICALL Java_com_jacob_com_DispatchEvents_init3
   
   BSTR typeLib = NULL;
   if (_typelib != NULL){
+  	  // why is this UTF instead of unicode? Then we could probably drop the A2W
 	  const char *typelib = env->GetStringUTFChars(_typelib, NULL);
 	  typeLib = A2W(typelib);
+	  // should we call env->ReleaseStringUTFChars(,) to free the memory like we do everywhere lese?
+	  
 	  //printf("we have a type lib %ls\n",typeLib);
   }
 
   // find progid if any
   LPOLESTR bsProgId = NULL;
-    if (_progid!=NULL) {
-    const char *progid = env->GetStringUTFChars(_progid, NULL);
+  if (_progid!=NULL) {
+    	// why is this UTF instead of unicode?  Then we could probably drop the A2W
+    	const char *progid = env->GetStringUTFChars(_progid, NULL);
 		bsProgId = A2W(progid);
+	  // should we call env->ReleaseStringUTFChars(,) to free the memory like we do everywhere lese?
 		//printf("we have an applicaton %ls\n",bsProgId);
   }
   
diff --git a/jni/SafeArray.cpp b/jni/SafeArray.cpp
index 050290f..661df53 100644
--- a/jni/SafeArray.cpp
+++ b/jni/SafeArray.cpp
@@ -558,24 +558,26 @@ JNIEXPORT void JNICALL Java_com_jacob_com_SafeArray_fromStringArray
     V_VT(&v) = VT_BSTR;
     for(int i=0;i<len;i++) {
       jstring s = (jstring)env->GetObjectArrayElement(a, i);
-      const char *str = env->GetStringUTFChars(s, NULL);
+      // jacob report 1224219 should use unicode and not UTF-8 (Variant modified in previous report
+      const jchar *str = env->GetStringChars(s, NULL); 
       CComBSTR bs(str);
       V_VT(&v) = VT_BSTR;
       V_BSTR(&v) = bs.Copy();
       long x = i;
       SafeArrayPutElement(psa,&x,&v);
-      env->ReleaseStringUTFChars(s, str);
+      env->ReleaseStringChars(s, str);
       VariantClear(&v);
     }
   } else if (vt == VT_BSTR) {
     for(int i=0;i<len;i++) {
       jstring s = (jstring)env->GetObjectArrayElement(a, i);
-      const char *str = env->GetStringUTFChars(s, NULL);
+      // jacob report 1224219 should use unicode and not UTF-8 (Variant modified in previous report
+      const jchar *str = env->GetStringChars(s, NULL);
       CComBSTR bs(str);
       BSTR bstr = bs.Detach();
       long x = i;
       SafeArrayPutElement(psa,&x,bstr);
-      env->ReleaseStringUTFChars(s, str);
+      env->ReleaseStringChars(s, str);
     }
   } else {
     ThrowComFail(env, "safearray cannot be assigned from string\n", 0);
@@ -1613,11 +1615,17 @@ JNIEXPORT jstring JNICALL Java_com_jacob_com_SafeArray_getString__I
     }
     BSTR bs = V_BSTR(&v);
     jstring js = env->NewString(bs, SysStringLen(bs));
+    // jacob report 1224219 
+	// SafeArrayGetElement memory must be freed http://www.canaimasoft.com/f90VB/OnlineManuals/Reference/TH_31.htm
+    VariantClear(&v);
     return js;
   } else if (vt == VT_BSTR) {
     BSTR bs = NULL;
     SafeArrayGetElement(psa, &idx, &bs);
     jstring js = env->NewString(bs, SysStringLen(bs));
+    // jacob report 1224219 
+	// SafeArrayGetElement memory must be freed http://www.canaimasoft.com/f90VB/OnlineManuals/Reference/TH_31.htm
+    if (bs) free(bs);
     return js;
   }
   ThrowComFail(env, "safearray cannot get string", 0);