#include <iostream>
#include <vector>
#include <algorithm>
using namespace std;
const int stala=1e5+10;
vector<int>w[stala];
vector<int>w2[stala];
vector<pair<int,int> >pom;
vector<int>result;
int nr[stala];
int postorder[stala];
int gl[stala];
int jump[stala][17];
int dp[stala];
int l=0;
void dfs(int k,int p)
{
jump[k][0]=p;
for(int i=0;i<(int)w[k].size();i++){
if(w[k][i]!=p){
gl[w[k][i]]=gl[k]+1;
dfs(w[k][i],k);
}
nr[w[k][i]]=w2[k][i];
}
l++;
postorder[k]=l;
}
void pre_lca(int ile)
{
for(int i=1;i<=16;i++){
for(int j=1;j<=ile;j++){
jump[j][i]=jump[jump[j][i-1]][i-1];
}
}
}
int lca(int x,int y)
{
if(gl[y]>gl[x]){
swap(x,y);
}
for(int i=16;i>=0;i--){
if(gl[jump[x][i]]>=gl[y]){
x=jump[x][i];
}
}
if(x==y){
return x;
}
for(int i=16;i>=0;i--){
if(jump[x][i]!=jump[y][i]){
x=jump[x][i];
y=jump[y][i];
}
}
return jump[x][0];
}
void func()
{
int all_lca=pom[0].second;
dp[pom[0].second]++;
for(int i=1;i<(int)pom.size();i++){
all_lca=lca(all_lca,pom[i].second);
dp[pom[i].second]++;
int x=lca(pom[i-1].second,pom[i].second);
dp[x]--;
}
dp[all_lca]--;
}
void final_dfs(int k,int p,int wsp)
{
for(int i=0;i<(int)w[k].size();i++){
if(w[k][i]!=p){
final_dfs(w[k][i],k,wsp);
dp[k]+=dp[w[k][i]];
}
}
if(dp[k]>=wsp){
result.push_back(nr[k]);
}
}
int main()
{
ios_base::sync_with_stdio(0);
cin.tie(0);
cout.tie(0);
int ile,m,k;
cin>>ile>>m>>k;
for(int i=1;i<ile;i++){
int a,b;
cin>>a>>b;
w[a].push_back(b);
w[b].push_back(a);
w2[a].push_back(i);
w2[b].push_back(i);
}
gl[1]=1;
dfs(1,1);
pre_lca(ile);
for(int i=1;i<=m;i++){
pom.clear();
int il;
cin>>il;
for(int j=1;j<=il;j++){
int a;
cin>>a;
pom.push_back({postorder[a],a});
}
sort(pom.begin(),pom.end());
func();
}
final_dfs(1,0,k);
sort(result.begin(),result.end());
cout<<(int)result.size()<<"\n";
for(int i=0;i<(int)result.size();i++){
cout<<result[i]<<" ";
}
return 0;
}
