#include <bits/stdc++.h>
#pragma GCC optimize("Ofast,unroll-loops")
using namespace std;
#define endl '\n'
#define ll long long
#define all(x) x.begin(), x.end()
struct T {
int depth, place;
};
bool operator < (T a, T b) {
return a.depth < b.depth;
}
priority_queue<T> pq;
int main() {
ios_base::sync_with_stdio(false);
cin.tie(NULL);
int n;
cin >> n;
vector<int> g[n + 2];
for (int i = 1; i <= n; i++) {
int sz; cin >> sz;
for (int j = 0; j < sz; j++) {
int t; cin >> t;
g[t].push_back(i);
}
}
ll ans = LLONG_MAX, mnDepth = LLONG_MAX;
for (int boss = 1; boss <= n; boss++) {
queue<int> q;
bool visited[n + 2];
int to[n + 2], depth[n + 2];
memset(depth, 0, sizeof(depth));
vector<int> leaf;
memset(visited, 0, sizeof(visited));
q.push(boss);
visited[boss] = 1;
int cnt = 0, mxDepth = 0;
while (q.size() > 0) {
cnt++;
int fr = q.front();
q.pop();
mxDepth += depth[fr];
bool next = 0;
for (auto x : g[fr]) {
if (visited[x]) continue;
depth[x] = depth[fr] + 1;
next = 1;
to[x] = fr;
visited[x] = 1;
q.push(x);
}
if (!next) leaf.push_back(fr);
}
if (mxDepth > mnDepth) continue;
if (cnt != n) continue;
int vals[n + 2];
memset(vals, 0, sizeof(vals));
for (auto x : leaf) {
pq.push({depth[x], x});
vals[x] = 1;
}
while (pq.size() > 0) {
int top = pq.top().place;
pq.pop();
if (vals[to[top]] == 0) vals[to[top]] = vals[top] + 1;
else vals[to[top]] += vals[top];
if (to[top] != boss && vals[to[top]] == vals[top] + 1) pq.push({depth[to[top]], to[top]});
}
cout << endl;
ll sum = 0;
for (int i = 1; i <= n; i++) {
sum += vals[i];
// cout << i << " : " << vals[i] << "\n";
}
// cout << boss << " " << sum << endl;
mnDepth = mxDepth;
ans = min(ans, sum);
}
cout << ans;
}
