解释两种方法内存消耗不同的原因

前：

 private static Dictionary<RuleKey, bool> mergeRules(
         Dictionary<RuleKey, bool> topPriorityRules, Dictionary<RuleKey, bool> secondaryRules)
      {
         var resolvedRules = topPriorityRules.Concat(secondaryRules.Where(kvp => !topPriorityRules.ContainsKey(kvp.Key)))
            .ToDictionary(x => x.Key, x => x.Value);
         return resolvedRules;
      }

后：

  private static Dictionary<TKey, TValue> merge<TKey, TValue>(
             Dictionary<TKey, TValue> topPriorityRules, Dictionary<TKey, TValue> secondaryRules)
          {
             var resolvedRules = topPriorityRules;

             foreach (var rule in secondaryRules)
                if(!resolvedRules.ContainsKey(rule.Key))
                   resolvedRules.Add(rule.Key, rule.Value);
             return resolvedRules;
          }

有2个权限集词典需要合并为一个。第一种情况，当方法被调用 500 次时。该程序占用的 RAM 量高达 1GB。在第二种情况下，大约 300MB。为什么内存消耗有这样的差异？毕竟，事实上，这些方法做同样的事情。