diff --git a/xml/chapter2/section3/subsection4.xml b/xml/chapter2/section3/subsection4.xml
index ebfffc325..bf176a3ab 100644
--- a/xml/chapter2/section3/subsection4.xml
+++ b/xml/chapter2/section3/subsection4.xml
@@ -235,12 +235,7 @@ BACADAEAFABBAAAGAH</TD></TR></TABLE>
     there is only one node left, which is the root of the entire tree.
     Here is how the Huffman tree of figure<SPACE/><REF NAME="fig:huffman"/> was
     generated:
-    <SPLIT>
-      <SCHEME>
-	<FIGURE src="img_tex/ch2_3_4-3.png"></FIGURE>
-      </SCHEME>
-      <JAVASCRIPT>
-	<TABLE>
+    <TABLE>
 	  <TR SINGLESPACE="yes">
             <TD>
 	      Initial leaves
@@ -305,9 +300,7 @@ BACADAEAFABBAAAGAH</TD></TR></TABLE>
 	      <LATEXINLINE>$\{$</LATEXINLINE>(<LATEXINLINE>$\{$</LATEXINLINE>A B C D E F G H<LATEXINLINE>$\}$</LATEXINLINE> 17)<LATEXINLINE>$\}$</LATEXINLINE>
 	    </TD>
 	  </TR>
-	</TABLE>
-      </JAVASCRIPT>
-    </SPLIT>
+    </TABLE>
     The algorithm does not always specify a unique tree, because there may
     not be unique smallest-weight nodes at each step.  Also, the choice of
     the order in which the two nodes are merged (i.e., which will be the
@@ -1052,12 +1045,7 @@ equal(sample_tree,
     <INDEX>rock songs, 1950s</INDEX>
     rock songs.  (Note that the <QUOTE>symbols</QUOTE> of an
     <QUOTE>alphabet</QUOTE> need not be individual letters.)
-    <SPLIT>
-      <SCHEME>
-	<FIGURE src="img_tex/ch2_3_4-4.png"></FIGURE>
-      </SCHEME>
-      <JAVASCRIPT>
-	<TABLE>
+    <TABLE>
 	  <TR SINGLESPACE="yes">
             <TD>
 	      A
@@ -1128,9 +1116,7 @@ equal(sample_tree,
 	      <SPACE/><SPACE/>1
 	    </TD>
 	  </TR>
-	</TABLE>
-      </JAVASCRIPT>
-    </SPLIT>
+    </TABLE>
     Use
     <SPLITINLINE>
       <SCHEME><SCHEMEINLINE>generate-huffman-tree</SCHEMEINLINE></SCHEME>
diff --git a/xml/chapter3/section1/subsection3.xml b/xml/chapter3/section1/subsection3.xml
index 101b341e2..f4a44a220 100644
--- a/xml/chapter3/section1/subsection3.xml
+++ b/xml/chapter3/section1/subsection3.xml
@@ -1045,6 +1045,7 @@ paul_acc("withdraw", "rosebud")(50); // Withdraws 50, should return 40
   </EXERCISE>
 
   <EXERCISE>
+    <LABEL NAME="ex:3_8"/>
     When we defined the evaluation model in
     section<SPACE/><REF NAME="sec:evaluating-combinations"/>, we said that the
     <INDEX>order of evaluation<SUBINDEX><ORDER>JavaScript</ORDER>in JavaScript</SUBINDEX></INDEX>
diff --git a/xml/chapter3/section2/section2.xml b/xml/chapter3/section2/section2.xml
index b69abbc1d..cb83d85d0 100644
--- a/xml/chapter3/section2/section2.xml
+++ b/xml/chapter3/section2/section2.xml
@@ -303,4 +303,7 @@
   <!-- Subsection 4 : Internal Definitions -->
   &amp;subsection3.2.4;
 
+  <!-- Subsection 5 : CSE Machine -->
+  &amp;subsection3.2.5;
+
 </SECTION>
diff --git a/xml/chapter3/section2/subsection5.xml b/xml/chapter3/section2/subsection5.xml
new file mode 100644
index 000000000..8e66d0500
--- /dev/null
+++ b/xml/chapter3/section2/subsection5.xml
@@ -0,0 +1,52 @@
+<SUBSECTION> 
+  <NAME>
+    CSE Machine
+  </NAME>
+
+  <LABEL NAME="sec:cse-machine"/>
+  <INDEX>CSE machine<OPEN/></INDEX>
+
+  <SPLIT>
+    <JAVASCRIPT>
+      <TEXT>
+	The environment model as presented so far focuses on how functions can refer
+	to their parameters, locally declared names, and names that are declared
+	outside the function. We achieve this by evaluating statements and expressions
+	with respect to a <EM>current environment</EM>. It does not specify how
+	we keep track of environments as computation proceeds. For example, when we
+	evaluate an expression <JAVASCRIPTINLINE>f(x) + y</JAVASCRIPTINLINE>, we
+	need to evaluate <JAVASCRIPTINLINE>x</JAVASCRIPTINLINE> in the current
+	environment, establish as the new current environment the environment of
+	<JAVASCRIPTINLINE>f</JAVASCRIPTINLINE> extended by a binding of its
+	parameter to the value of <JAVASCRIPTINLINE>x</JAVASCRIPTINLINE>, and
+	evaluate the body of <JAVASCRIPTINLINE>f</JAVASCRIPTINLINE> in this
+	extended environment. But what environment should we use for evaluating
+	<JAVASCRIPTINLINE>y</JAVASCRIPTINLINE> after
+	<JAVASCRIPTINLINE>f</JAVASCRIPTINLINE> returns?
+	In this section, we extend the 
+      </TEXT>
+
+      <SUBHEADING><NAME>Evaluating arithmetic expressions</NAME></SUBHEADING>
+      
+      <TEXT>
+	Exercise<SPACE/><REF NAME="ex:3_8"/> shows that the presence of
+	assignments makes the result of a program depend on the order in which
+	the operands of an operator combination are evaluated. To remove
+	ambiguities that arise from this, the JavaScript standard specifies
+	left-to-right evaluation of operands.
+
+	As an example, consider the evaluation of the arithmetic expression statement
+	<SNIPPET>
+	  <JAVASCRIPT>
+1 + (2 * 3);
+	  </JAVASCRIPT>
+	</SNIPPET>
+	The expression is decomposed into its operands
+	<JAVASCRIPTINLINE>1</JAVASCRIPTINLINE> and
+	<JAVASCRIPTINLINE>2 * 3</JAVASCRIPTINLINE>, followed by the
+	<EM>instruction</EM> to add their results.
+      </TEXT>
+  <INDEX>CSE machine<CLOSE/></INDEX>
+    </JAVASCRIPT>
+  </SPLIT>
+</SUBSECTION>
diff --git a/xml/chapter3/section5/subsection2.xml b/xml/chapter3/section5/subsection2.xml
index 26eb77fde..505fa44b0 100644
--- a/xml/chapter3/section5/subsection2.xml
+++ b/xml/chapter3/section5/subsection2.xml
@@ -578,7 +578,15 @@ const fibs = pair(0,
     by adding <SCHEMEINLINE>fibs</SCHEMEINLINE> to itself shifted by one place:
     <SPLIT>
       <SCHEME>
-        <FIGURE src="img_tex/ch3_5_2-1.png"></FIGURE>
+	<LATEX>
+\[
+\begin{array}{ccccccccccccl}
+  &amp;   &amp; 1 &amp; 1 &amp; 2 &amp; 3 &amp; 5 &amp;  8 &amp; 13 &amp; 21 &amp; \ldots &amp; = &amp; \texttt{(stream-cdr fibs)} \\
+  &amp;   &amp; 0 &amp; 1 &amp; 1 &amp; 2 &amp; 3 &amp;  5 &amp;  8 &amp; 13 &amp; \ldots &amp; = &amp; \texttt{fibs} \\ \hline
+0 &amp; 1 &amp; 1 &amp; 2 &amp; 3 &amp; 5 &amp; 8 &amp; 13 &amp; 21 &amp; 34 &amp; \ldots &amp; = &amp; \texttt{fibs}
+\end{array}
+\]
+	</LATEX>
       </SCHEME>
       <JAVASCRIPT>
 	<LATEX>
diff --git a/xml/chapter5/section5/subsection1.xml b/xml/chapter5/section5/subsection1.xml
index 643606844..161b5da75 100644
--- a/xml/chapter5/section5/subsection1.xml
+++ b/xml/chapter5/section5/subsection1.xml
@@ -514,7 +514,16 @@ go_to(reg("continue"))
   <!--  %Whew!  What a pain it was to set this, because of the blank lines. -->
   <SPLIT>
     <SCHEME>
-      <FIGURE src="img_original/preserving_table.svg"></FIGURE>
+      \[
+      \begin{array}{l|l|l|l}
+        \textit{seq}_1 &amp; \texttt{(save}\ \textit{reg}_1\texttt{)}    &amp; \texttt{(save}\ \textit{reg}_2\texttt{)}     &amp; \texttt{(save}\ \textit{reg}_2\texttt{)}    \\
+        \textit{seq}_2 &amp; \textit{seq}_1                             &amp; \textit{seq}_1                              &amp; \texttt{(save}\ \textit{reg}_1\texttt{)}    \\
+                       &amp; \texttt{(restore}\ \textit{reg}_1\texttt{)} &amp; \texttt{(restore}\ \textit{reg}_2\texttt{)}  &amp; \textit{seq}_1                             \\
+                       &amp; \textit{seq}_2                             &amp; \textit{seq}_2                              &amp; \texttt{(restore}\ \textit{reg}_1\texttt{)} \\
+                       &amp;                                            &amp;                                             &amp; \texttt{(restore}\ \textit{reg}_2\texttt{)} \\
+                       &amp;                                            &amp;                                             &amp; \textit{seq}_2
+      \end{array}
+      \]
     </SCHEME>
     <JAVASCRIPT>
     <LATEX>
@@ -525,7 +534,7 @@ go_to(reg("continue"))
                        &amp; \texttt{restore(}\textit{reg}_1\texttt{),} &amp; \texttt{restore(}\textit{reg}_2\texttt{),}  &amp; \textit{seq}_1                             \\
                        &amp; \textit{seq}_2                             &amp; \textit{seq}_2                              &amp; \texttt{restore(}\textit{reg}_1\texttt{),} \\
                        &amp;                                            &amp;                                             &amp; \texttt{restore(}\textit{reg}_2\texttt{),} \\
-                       &amp;                                            &amp;                                             &amp; \textit{seq}_2 
+                       &amp;                                            &amp;                                             &amp; \textit{seq}_2
       \end{array}
       \]
     </LATEX>