def dfs(graph, visited, here, visited_vs):

    visited[here]=True

    neighbors=[i for i in range(26) if graph[here][i]==1]

    for neighbor in neighbors:
        if visited[neighbor]==False:
            visited, visited_vs=dfs(graph, visited, neighbor, visited_vs)

    visited_vs.append(here)

    return visited, visited_vs

def dfsAll(graph):

    visited=[False for i in range(26)] # 어차피 알파벳

    temp=[]
    for v in range(26):
        if visited[v]==False:
            visited_vs=[]
            visited, visited_vs=dfs(graph, visited, v, visited_vs)

            # visited_vs.reverse() # reverse를 해야 topological sorting이 가능? 
            temp+=visited_vs

    temp.reverse() # 먼저 순서로 만들어 주기 위해 다시 뒤집기?

    answer=''
    for t in temp:
        answer+=chr(t+97)

    return answer

def getAnswer(words):
    graph=[[0 for i in range(26)] for i in range(26)] # 그래프의 인접행렬표현
    # 'a':0, 'b':1 ... 하려면... ord(character)-ord('a')

    # Make graph
    for w in range(len(words)-1):
        word1=words[w]
        word2=words[w+1]

        for i in range(len(word1)): # 앞 단어가 끝날때까지만 보면 됨
            c1=word1[i]
            c2=word2[i]
            if c1!=c2:
                graph[ord(c1)-ord('a')][ord(c2)-ord('a')]=1
                break

    # Check cycle
    for i in range(26):
        for j in range(i, 26):
            if graph[i][j]+graph[j][i]==2:
                return "INVALID HYPOTHESIS"

    # Make topological sorting
    answer=dfsAll(graph)
    return answer

if __name__=="__main__":
    n=int(input())

    answers=[]

    for i in range(n):
        nbwords=int(input())
        words=[]
        for j in range(nbwords):
            words.append(input())

        answer=getAnswer(words)
        answers.append(answer)

    for answer in answers:
        print(answer)