The Gemara states that the list Haman's ten sons should be read in one breath. The common practice is to also read the word "Aseret" (ten), which sums up the list, in the same breath as the names. The reason for this custom is to emphasize that all ten sons died simultaneously, even though they were hung one after the other.
The custom is to also read the words that proceed their names, "V'et Hamesh M'ot Ish"-(and five hundred men) in the same breath, as well. The reason is that these men comprised Haman's sons' army; each son was captain over fifty men.
The Shulhan Aruch (siman 690) records this Halacha with the words, "V'sarich Lomar…"- (And it MUST be read…). The Poskim debate whether Maran's language indicates that this Halacha is critical to fulfilling the obligation. That is, if one did not read the names in this manner, B'diavad-after the fact, does he have to go back and read again? In any event, Maran, in his Bet Yosef, cites the opinion of Tosafot that this Halacha is L'Hathila, but if one did not do so, he has fulfilled the Misva after the fact. From this, it can be derived that Maran's intention in the Shulhan Aruch is also only L'Hathila. This is clearly the Rema's opinion, as well.
------
Maran, in Siman 690, records another Halacha that if the reader made mistakes in pronunciation, it is not necessary to be meticulous to correct him. The Megillah does not have the status of a "Sefer"-a holy book, but of an "Iggeret"-a letter. Maran does not distinguish between levels of severity of mistakes. Thus, it seems that even if the mistake changes the meaning of the word, it Is not necessary to correct the reader.
However, Maran continues and presents a "Yesh Omrim"-an additional opinion, that this leniency only applies to mistakes that do not change the meaning of the word. This is based on the incident, recorded in the Talmud Yerushalmi, in which the sages did not correct the reader who made a minor mistake in pronouncing the word "Yehudim" (Jews). This implies that the leniency only applies to mistakes that do not change the meaning. Interestingly, Hacham Ovadia rules in accordance with the "Yesh," additional opinion, that any mistake that changes the meaning of the word is critical, and the reader must go back.
--------
The Bet Yosef (Siman 690) cites the incredible opinion of Rabbenu Avraham ibn Ezra (1089-1167, Spain). He holds that the Megillah was originally written without "Sof Pasuk"-periods, i.e. end of sentence punctuation. Since it was given as an "Iggeret"- a letter, it was not formally broken up into verses. Only later, Ezra HaSofer added sentence punctuation to the Megillah. Therefore, the Ibn Ezra holds that the Megillah should be read in one long flow, without stopping at the end of the Pesukim-verses.
Many authorities question this opinion. Rav Shlomo Alkabes (1500-1576, Tsfat), in his Manot Levi commentary on the Megillah, argues that if the punctuation added by Ezra was not to be applied when reading the Megillah, for what purpose did he add it? On the contrary, the fact that Ezra added periods indicates that they are important and must be followed. This is how Hacham Ovadia rules. The accepted practice is to allow the reader to take a breath in between each Pasuk.
SUMMARY
If the reader did not read the list of Haman's sons in one breath, he does not have to go back and read it again.
Reading mistakes that do not change the meaning of the word are not critical, but mistakes that do change the meaning are critical, and the reader must go back and read correctly.
The reader should conform to the end of verse punctuation, and he can take a breath in between each verse.
