Reading:
ntheorem with the [amsthm] option ignores styling of a theorem environment
I adopted the suggested proof environment for ntheorem without the amsmath option. I also made sure to add the thmmarks options to ntheorem, to enable end-of-theorem/end-of-proof marks. However - if I write this document:
\documentclass{article}
\makeatletter
\usepackage[thmmarks]{ntheorem}
\usepackage{amssymb} % for \blacksquare
\usepackage{mathtools}
\newcommand{\proofname}{Proof}
\newcounter{proof}\newcounter{currproofctr}\newcounter{endproofctr}%
\newenvironment{proof}[1][\proofname]{
\th@nonumberplain
\def\theorem@headerfont{\itshape}%
\normalfont
@thm{proof}{proof}{{#1}.}}%
{@endtheorem} % and this ends the definition of the proof environment
\newcommand\proofSymbol{\ensuremath{\blacksquare}}
\qedsymbol{\ensuremath{\blacksquare}}
\makeatother
\begin{document}
\begin{proof}
The quick brown fox jumped over the lazy dog 1.
\end{proof}
\begin{proof}
The quick brown fox jumped over the lazy dog 2.
\begin{align}
42
\end{align}
\end{proof}
and some text afterwards.
\end{document}
I get a black square on the first proof, but not the second. Why is that?

\openboxpart is not used I think? You redefine\proofSymbolat the end of the preamble, which discards the earlier\gdef. That is not the core issue of course (the output is the same after deleting the\openboxdefinition). – Marijn Jan 17 '21 at 12:09ntheorempackage rather than theamsthmpackage? The reason I ask is that the latter package defines a very handyproofenvironment and even provides the macro\qedhereto place the QED symbol in some other location than at the very end. – Mico Jan 17 '21 at 12:11\newenvironment{proof}, with a regular\theoremsymbol{\ensuremath{\blacksquare}}\newtheorem{myproof}{Proof}you see the same behavior (first proof with square, second without square). Thentheoremmanual mentions (although not very clearly) that for theorems ending with math you needamsthmfor end marks (page 15), so it you want to use this withoutamsthmthen you probably need to lift some code from that package as well. – Marijn Jan 17 '21 at 12:30\openbox. Do you have any idea what code specifically I would need to lift, or shall I just dig into that to look? – einpoklum Jan 17 '21 at 12:31amsthmuses over 150 lines of code for just the QED symbol, so I gues the other way around (i.e., makentheoremwith theamsthmoption respect formatting modifications) would be easier to do. – Marijn Jan 17 '21 at 14:31amsthmmay be problematic for use with Hebrew. – einpoklum Jan 17 '21 at 17:18amsthm. – Marijn Jan 17 '21 at 18:28