#include<bits/stdc++.h>
#define INF 1e18
#define fi first
#define se second
#define FOR(i, k, n) for(ll i = k; i <= n; i++)
#define FOR1(i, k, n) for(ll i = k; i >= n; i--)
#define pb push_back
#define fastio ios::sync_with_stdio(0); cin.tie(0); cout.tie(0)
#define vi vector<int>
#define pii pair<int, int>
#define vii vector<pii>
#define ll long long
#define vll vector<ll>
#define pll pair<ll, ll>
#define re return 0
#define mii map<int, int>
#define input "CANDY.inp"
#define output "CANDY.out"
#define rf freopen(input, "r", stdin); freopen(output, "w", stdout)
#define int ll
using namespace std;
const int maxn = 1e5 + 5;
const int mod = 1e9 + 7;
const int base = 998244353;
void add(int &a, int b)
{
a += b;
if(a >= mod) a -= mod;
if(a < 0) a += mod;
}
mt19937 rd(chrono::steady_clock::now().time_since_epoch().count());
int rand(int l, int r)
{
return uniform_int_distribution<int>(l, r) (rd);
}
int n, m, q;
int depth[maxn], version[maxn], cntnode, cnt[20][maxn], par[20][maxn];
vi query[maxn], adj[maxn], vv;
struct shape{
int cnt, l, r;
ll sum;
shape(){
l = r = cnt = sum = 0;
}
} T[maxn * 25];
int build(int l, int r)
{
int node = ++cntnode;
if(l == r)
{
return node;
}
int mid = (l + r) >> 1;
T[node].l = build(l, mid);
T[node].r = build(mid + 1, r);
return node;
}
void DFS(int u, int p)
{
for(auto v : adj[u])
{
if(v == p)
continue;
depth[v] = depth[u] + 1;
DFS(v, u);
}
}
int update(int l, int r, int pos, int old)
{
int node = ++cntnode;
T[node] = T[old];
if(l == r)
{
T[node].cnt++;
T[node].sum += vv[l];
return node;
}
int mid = (l + r) >> 1;
if(pos <= mid)
T[node].l = update(l, mid, pos, T[old].l);
else
T[node].r = update(mid + 1, r, pos, T[old].r);
T[node].cnt = T[T[node].l].cnt + T[T[node].r].cnt;
T[node].sum = T[T[node].l].sum + T[T[node].r].sum;
return node;
}
void DFS1(int u)
{
for(auto v : adj[u])
{
if(v == par[0][u])
continue;
par[0][v] = u;
FOR(i, 1, 18)
{
par[i][v] = par[i - 1][par[i - 1][v]];
cnt[i][v] = cnt[i - 1][v] + cnt[i - 1][par[i - 1][v]];
}
if(query[v].empty())
version[v] = version[u];
else
{
int id = lower_bound(vv.begin(), vv.end(), query[v][0]) - vv.begin();
version[v] = update(1, vv.size(), id, version[u]);
FOR(i, 1, query[v].size() - 1)
{
id = lower_bound(vv.begin(), vv.end(), query[v][i]) - vv.begin();
version[v] = update(1, vv.size(), id, version[v]);
}
}
DFS1(v);
}
}
pii LCA(int u, int v)
{
if(depth[u] < depth[v])
swap(u, v);
int h = depth[u] - depth[v];
int dem = 0;
FOR1(i, 18, 0)
if(h >= (1 << i))
{
dem += cnt[i][u];
u = par[i][u];
h -= (1 << i);
}
if(u == v)
return {u, dem};
FOR1(i, 18, 0)
if(par[i][v] != par[i][u])
{
dem += cnt[i][v];
dem += cnt[i][u];
u = par[i][u];
v = par[i][v];
}
dem += cnt[0][u] + cnt[0][v];
return {par[0][u], dem};
}
pll tinh(int x, int y, int z)
{
int dem = 0;
ll sum = 0;
dem = T[x].cnt + T[y].cnt - 2 * T[z].cnt;
sum = T[x].sum + T[y].sum - 2 * T[z].sum;
return {dem, sum};
}
int get(int l, int r, int u, int v, int lca, ll val)
{
pll tmp = tinh(u, v, lca);
if(tmp.se <= val)
return tmp.fi;
if(l == r)
{
pll tmp = tinh(u, v, lca);
int dem = tmp.fi;
return min((ll)dem, val / vv[l]);
}
int mid = (l + r) >> 1;
tmp = tinh(T[u].l, T[v].l, T[lca].l);
if(tmp.se <= val)
{
return tmp.fi + get(mid + 1, r, T[u].r, T[v].r, T[lca].r, val - tmp.se);
}
return get(l, mid, T[u].l, T[v].l, T[lca].l, val);
}
signed main()
{
fastio;
vii adj1;
cin >> n >> m >> q;
FOR(i, 1, n - 1)
{
int u, v;
cin >> u >> v;
adj[u].pb(v);
adj[v].pb(u);
adj1.pb({u, v});
}
DFS(1, 0);
FOR(i, 1, m)
{
int id, w;
cin >> id >> w;
vv.pb(w);
id--;
int u = adj1[id].fi, v = adj1[id].se;
if(depth[u] < depth[v])
swap(u, v);
cnt[0][u]++;
query[u].pb(w);
}
vv.pb(0);
sort(vv.begin(), vv.end());
vv.erase(unique(vv.begin(), vv.end()), vv.end());
version[1] = build(1, vv.size());
DFS1(1);
while(q--)
{
int u, v, x;
ll y;
cin >> u >> v >> x >> y;
pii tmp = LCA(u, v);
int dem = tmp.se, lca = tmp.fi;
int ans = x - dem;
ans += get(1, vv.size(), version[u], version[v], version[lca], y);
ans = max(ans, -1ll);
cout << ans << "\n";
}
re;
}
