diff --git a/src/math/lp/nex.cpp b/src/math/lp/nex.cpp
index 8481f5ddc..6d5abe4f9 100644
--- a/src/math/lp/nex.cpp
+++ b/src/math/lp/nex.cpp
@@ -83,18 +83,33 @@ void promote_children_of_sum(ptr_vector<nex> & children, nex_lt lt ) {
     } 
 }
 
-void promote_children_of_mul(vector<nex_pow> & children, nex_lt lt) {
+bool eat_scalar(nex_scalar *& r, nex_pow& p) {
+    if (!p.e()->is_scalar())
+        return false;
+    nex_scalar *pe = to_scalar(p.e());
+    if (r == nullptr) {
+        r = pe;
+        r->value() = r->value().expt(p.pow());                
+    } else {
+        r->value() *= pe->value().expt(p.pow());
+    }
+    return true;
+}
+
+void simplify_children_of_mul(vector<nex_pow> & children, nex_lt lt) {
+    nex_scalar* r = nullptr;
     TRACE("nla_cn_details", print_vector(children, tout););
     vector<nex_pow> to_promote;
     int skipped = 0;
-    for(unsigned j = 0; j < children.size(); j++) {
+    for(unsigned j = 0; j < children.size(); j++) {        
         nex_pow& p = children[j];
+        if (eat_scalar(r, p)) {
+            skipped++;
+            continue;
+        }
         (p.e())->simplify(p.ee(), lt);
         if ((p.e())->is_mul()) {
             to_promote.push_back(p);
-        } else if (ignored_child(p.e(), expr_type::MUL)) {
-            skipped ++;
-            continue;
         } else {
             unsigned offset = to_promote.size() + skipped;
             if (offset) {
@@ -104,14 +119,18 @@ void promote_children_of_mul(vector<nex_pow> & children, nex_lt lt) {
     }
     
     children.shrink(children.size() - to_promote.size() - skipped);
-    
+
     for (nex_pow & p : to_promote) {
         for (nex_pow& pp : to_mul(p.e())->children()) {
-            SASSERT(!ignored_child(pp.e(), expr_type::MUL));
-            children.push_back(nex_pow(pp.e(), pp.pow() * p.pow()));            
+            if (!eat_scalar(r, pp))
+                children.push_back(nex_pow(pp.e(), pp.pow() * p.pow()));            
         }
     }
 
+    if (r != nullptr) {
+        children.push_back(nex_pow(r));
+    }
+        
     mul_to_powers(children, lt);
     
     TRACE("nla_cn_details", print_vector(children, tout););
diff --git a/src/math/lp/nex.h b/src/math/lp/nex.h
index bb9a256f5..c874c7627 100644
--- a/src/math/lp/nex.h
+++ b/src/math/lp/nex.h
@@ -108,7 +108,7 @@ public:
     lpvar var() const {  return m_j; }
     lpvar& var() {  return m_j; } // the setter
     std::ostream & print(std::ostream& out) const {
-        out << 'v' <<  m_j;
+        out <<  (char)('a' + m_j);
         return out;
     }    
 
@@ -141,7 +141,7 @@ const nex_sum* to_sum(const nex*a);
 
 void promote_children_of_sum(ptr_vector<nex> & children, nex_lt);
 class nex_pow;
-void promote_children_of_mul(vector<nex_pow> & children, nex_lt);
+void simplify_children_of_mul(vector<nex_pow> & children, nex_lt);
 
 class nex_pow {
     nex* m_e;
@@ -243,7 +243,7 @@ public:
         TRACE("nla_cn_details", tout << "**e = " << **e << "\n";);
         *e = this;
         TRACE("nla_cn_details", tout << *this << "\n";);
-        promote_children_of_mul(m_children, lt);
+        simplify_children_of_mul(m_children, lt);
         if (size() == 1 && m_children[0].pow() == 1) 
             *e = m_children[0].e();
         TRACE("nla_cn_details", tout << *this << "\n";);
@@ -424,6 +424,11 @@ inline const nex_scalar* to_scalar(const nex*a)  {
     return static_cast<const nex_scalar*>(a);
 }
 
+inline nex_scalar* to_scalar(nex*a)  {
+    SASSERT(a->is_scalar());
+    return static_cast<nex_scalar*>(a);
+}
+
 inline const nex_mul* to_mul(const nex*a) {
     SASSERT(a->is_mul());
     return static_cast<const nex_mul*>(a);
diff --git a/src/test/lp/lp.cpp b/src/test/lp/lp.cpp
index 129918020..1c5ab4a89 100644
--- a/src/test/lp/lp.cpp
+++ b/src/test/lp/lp.cpp
@@ -85,19 +85,23 @@ void test_simplify() {
         );
     enable_trace("nla_cn");
     enable_trace("nla_cn_details");
-    nex_var* a = cn.get_nex_creator().mk_var(0);
-    nex_var* b = cn.get_nex_creator().mk_var(1);
-    nex_var* c = cn.get_nex_creator().mk_var(2);
-    auto & r = cn.get_nex_creator();
-    auto m = r.mk_mul(); m->add_child_in_power(c, 2);
+    auto & creator = cn.get_nex_creator();
+    nex_var* a = creator.mk_var(0);
+    nex_var* b = creator.mk_var(1);
+    nex_var* c = creator.mk_var(2);
+    auto m = creator.mk_mul(); m->add_child_in_power(c, 2);
     TRACE("nla_cn", tout << "m = " << *m << "\n";); 
-    auto n = r.mk_mul(a);
+    auto n = creator.mk_mul(a);
     n->add_child_in_power(b, 7);
+    n->add_child(creator.mk_scalar(rational(3)));
+    n->add_child_in_power(creator.mk_scalar(rational(4)), 2);
+    n->add_child(creator.mk_scalar(rational(1)));
     TRACE("nla_cn", tout << "n = " << *n << "\n";); 
     m->add_child_in_power(n, 3);
+    n->add_child_in_power(creator.mk_scalar(rational(1, 3)), 2);
     TRACE("nla_cn", tout << "m = " << *m << "\n";); 
     
-    nex * e = r.mk_sum(a, r.mk_sum(b, m));
+    nex * e = creator.mk_sum(a, creator.mk_sum(b, m));
     TRACE("nla_cn", tout << "e = " << *e << "\n";);
     e->simplify(&e);
     TRACE("nla_cn", tout << "simplified e = " << *e << "\n";);