We may rewrite (?) as
\dim _{\mathbb{C}}\mathcal{A}_{m}^{\nu }=\frac{2m+n+2\nu }{n!(n-1)!}
\prod\limits_{j=1}^{n-1}(m+j)(m+2\nu +j) .
Setting r=r\left( m\right) =m+\nu +\frac{n}{2}, the product in (Proof) also reads
\mathcal{\wp }=\prod\limits_{j=1}^{n-1}\left( r-\frac{n}{2}-\nu +j\right) \left( r-\frac{n}{
2}+\nu +j\right) .
We treat the cases of n odd and n even separately as they correspond to different choices of theta functions.
\left( i\right) Case odd n\geq 0. The product (Proof) can also be presented as
\mathcal{\wp }=\prod\limits_{l=\frac{1}{2}+\nu }^{\frac{n}{2}+\nu -1}\left(
r^{2}-l^{2}\right) \prod\limits_{s=\frac{1}{2}-\nu }^{\frac{n}{2}-\nu
-1}\left( r^{2}-s^{2}\right)
and further it can be decomposed as
\mathcal{\wp }=\sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}r^{2p}.
where the numbers \gamma _{0}^{(\nu ,n)},...,\gamma _{n-1}^{(\nu ,n)} can be computed explicitly (see Sect.?). By using (?), the multiplicity in (Proof) takes the form
\dim _{\mathbb{C}}\mathcal{A}_{m}^{\nu }=\frac{2}{n!(n-1)!}
\sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}\left( m+\nu +\frac{n}{2}
\right) ^{2p+1}
which we may use to rewrite the trace formula (?) as
\begin{aligned}
\begin{split}
Tr\left( e^{\frac{1}{4}t\Delta _{\nu }}\right) &=e^{\left( \frac{n^{2}
}{4}+\nu ^{2}\right) t}\sum\limits_{m=0}^{+\infty }\left[ \frac{2}{n!(n-1)!}
\sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}\left( m+\nu +\frac{n}{2}
\right) ^{2p+1}\right] e^{-\left( m+\frac{n}{2}+\nu \right) ^{2}t} \\
&=\frac{e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{n!(n-1)!}\mathcal{S}
\end{split}
\end{aligned}
with
\mathcal{S}:=\sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}\left(
2\sum\limits_{m=0}^{+\infty }\left( m+\nu +\frac{n}{2}\right)
^{2p+1}e^{-\left( m+\frac{n}{2}+\nu \right) ^{2}t}\right) .
The latter sum can be cast into two parts as
\mathcal{S}=2\sum\limits_{\mu =\nu +\frac{1}{2}}^{+\infty }\mu e^{-\mu
^{2}t}\sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}\mu
^{2p}-2\sum\limits_{\mu =\nu +\frac{1}{2}}^{\nu +\frac{n}{2}-1}\mu e^{-\mu
^{2}t}\sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}\mu ^{2p}
Recalling (?)-(?), we see that for \mu raining over the set \{\nu +\frac{1}{2},\nu +\frac{3}{2},\cdots ,\nu +\frac{n}{2}-1\} the sum
\sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}\mu ^{2p}=0
and therefore the sum (?) reduces to
\mathcal{S}=2\sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}\mathcal{R}
_{p}^{\left( \nu \right) }\left( t\right)
where
\mathcal{R}_{p}^{\left( \nu \right) }\left( t\right) =\sum\limits_{\mu =\nu +
\frac{1}{2}}^{+\infty }\mu ^{2p+1}e^{-\mu ^{2}t}
By setting \mu =j+\frac{1}{2}, we successively obtain
\begin{aligned}
\mathcal{R}_{p}^{\left( \nu \right) }\left( t\right) &=\sum\limits_{j=\nu
}^{+\infty }\left( j+\frac{1}{2}\right) ^{2p+1}e^{-\left( j+\frac{1}{2}
\right) ^{2}t} \\
&=\sum\limits_{j=0}^{+\infty }\left( j+\frac{1}{2}\right) ^{2p+1}e^{-\left( j+
\frac{1}{2}\right) ^{2}t}-\sum\limits_{j=0}^{\nu -1}\left( j+\frac{1}{2}
\right) ^{2p+1}e^{-\left( j+\frac{1}{2}\right) ^{2}t} \\
&=\frac{1}{2}\left( -\frac{d}{dt}\right) ^{p}\left[ \sum\limits_{j=0}^{+
\infty }\left( 2j+1\right) e^{-\left( j+\frac{1}{2}\right) ^{2}t}\right]
-\sum\limits_{j=0}^{\nu -1}\left( j+\frac{1}{2}\right) ^{2p+1}e^{-\left( j+
\frac{1}{2}\right) ^{2}t} \\
&=\frac{1}{2}\left( -\frac{d}{dt}\right) ^{p}\left[ \vartheta _{2}(t)\right]
-\sum\limits_{j=0}^{\nu -1}\left( j+\frac{1}{2}\right) ^{2p+1}e^{-\left( j+
\frac{1}{2}\right) ^{2}t}
\end{aligned}
where \vartheta _{2}(t) denotes the Jacobi’s theta function given by [M_1928, p.280]:
\vartheta _{2}(t)=\sum\limits_{j=0}^{+\infty }\left( 2j+1\right) e^{-\left(
j+\frac{1}{2}\right) ^{2}t}
while the finite sum in (?) may be expressed by the series
\sum\limits_{\ell =0}^{+\infty }\sigma _{p}^{(\nu )}\left( \ell \right) \
t^{\ell }
with coefficients
\sigma _{p}^{\left( \nu \right) }\left( \ell \right) =\frac{(-1)\ell }{\ell !
}\sum\limits_{j=0}^{\nu -1}\left( j+\frac{1}{2}\right) ^{2(p+\ell )+1}.
Now, we use the formula [prud1, p.597]:
\sum\limits_{k=0}^{m}(k+a)^q=\frac{1}{q+1}\left[ B_{q+1}(m+1+a)-B_{q+1}(a)\right], \quad q=1,2,\cdots ,
where B_n(z) are the Bernoulli polynomials, for m=\nu-1, q=2(\ell+p)+1, a=\frac{1}{2}, to get
\sigma _{p}^{\left( \nu \right) }\left( \ell \right) =\frac{(-1)^{\ell}}{2(p+\ell +1)\ell !}\left[ B_{2(p+\ell +1)}\left(\nu+\frac{1}{2}\right)-B_{2(p+\ell +1)}\left(\frac{1}{2}\right)\right].
Summarizing the above calculations, then the sum in (?) also reads
\mathcal{S}=\sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}\left[ \left( -
\frac{d}{dt}\right) ^{p}\vartheta _{2}(t)+\sum\limits_{\ell =0}^{+\infty}\frac{(-1)^{\ell}}{(p+\ell +1)\ell !}\left[ B_{2(p+\ell +1)}\left(\frac{1}{2}\right)-B_{2(p+\ell +1)}\left(\nu+\frac{1}{2}\right)\right] \ t^{\ell }\right]
Applying now the asymptotic of the higher order derivative of \vartheta _{2}(t) as t\searrow 0^{+} ([M_1928, p.281]) :
\left( \frac{d}{dt}\right) ^{p}\vartheta _{2}(t)\simeq \frac{(-1)^{p}p!}{
t^{1+p}}+\sum\limits_{s=0}^{+\infty }\frac{B_{s+p}t^{s}}{s!}
in terms of Bernoulli numbers \left( B_{d\text{\ }}\right) defined by the recursion relation [Aw2019, p.8]:
B_{d}=\frac{\left( -1\right) ^{d}}{d+1}\left( 1-2^{-2d-1}\right)B_{2d+2} ,
\text{ \ }d=0,1,2,...
By combining (?) and the formula [prud1, p.777]
B_d\left(\frac{1}{2}\right) =-(1-2^{1-d}) B_d,
we deduce that
B_{2(d+1)}\left(\frac{1}{2}\right) =(-1)^{d+1}(d+1)B_d.
For d=p+s (?) reads
B_{2(p+s+1)}\left(\frac{1}{2}\right) =(-1)^{p+s+1}(p+s+1)B_{p+s}.
Substituting (?) and (?) into (?), we get after simplifications
\mathcal{S} \simeq \sum\limits_{p=0}^{n-1}\gamma _{p}^{(\nu ,n)}\left[
p!t^{-1-p}+\sum\limits_{s=0}^{+\infty }\frac{(-1)^{s+1}B_{2(p+s+1)}\left(\nu+\frac{1}{2}\right)}{(p+s+1)s!}t^{s}\right] .
Therefore, the heat trace asymptotic formula becomes
\begin{split}
Tr\left( e^{\frac{1}{4}t\Delta _{\nu }}\right) &\simeq \frac{e^{\left( \frac{
n^{2}}{4}+\nu ^{2}\right) t}}{n!(n-1)!}\sum\limits_{p=0}^{n-1}\gamma
_{p}^{(\nu ,n)}\left[ \frac{p!}{t^{1+p}}+\sum\limits_{s=0}^{+\infty }
\frac{(-1)^{s+1}B_{2(p+s+1)}\left(\nu+\frac{1}{2}\right)}{(p+s+1)s!}
t^{s}\right] \\
&= \frac{e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{n!(n-1)!}\left[ \sum\limits_{p=0}^{n-1}\gamma
_{p}^{(\nu ,n)}\frac{p!}{t^{1+p}}+\sum\limits_{p=0}^{n-1}\gamma
_{p}^{(\nu ,n)}\sum\limits_{s=0}^{+\infty }
\frac{(-1)^{s+1}B_{2(p+s+1)}\left(\nu+\frac{1}{2}\right)}{(p+s+1)s!}
t^{s}\right] \\
& =\frac{\omega _{n}e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{(4\pi t)^{n}}
\left[ 1+\sum\limits_{p=0}^{n-2}\frac{\gamma _{p}^{(\nu ,n)}p!}{(n-1)!}
t^{n-p-1}+\sum\limits_{p=0}^{n-1}\frac{\gamma _{p}^{(\nu ,n)}}{(n-1)!}\sum\limits_{s=p}^{+\infty }\frac{(-1)^{s-p+1}B_{2(s+1)}\left(\nu+\frac{1}{2}\right)}{(s+1)(s-p)!} t^{n+s-p}\right]
\end{split}
The first sum in (?) can also be written as
\sum\limits_{d=0}^{n-1}\frac{\gamma _{n-1-d}^{(\nu ,n)}}{(n-1)!}\left(
n-1-d\right) !t^{d}
while the second one also reads
\sum\limits_{d=n}^{+\infty }\left( \sum\limits_{p=0}^{n-1}\frac{(-1)^{d-n+1}\gamma_{p}^{(\nu ,n)} B_{2(d-n+p+1)}\left(\nu+\frac{1}{2}\right)}{(n-1)!(d-n+p+1)(d-n)!} \right) t^{d} .
By introducing
\quad \Omega _{p}^{\left( \nu \right) }\left( k\right) :=\frac{(-1)^{k+1} B_{2(k+p+1)}\left(\nu+\frac{1}{2}\right)}{(k+p+1)k!} ,\qquad k=0,1,2,...,
we may rewrite (?) as
Tr\left( e^{\frac{1}{4}t\Delta _{\nu }}\right) \simeq \frac{\omega
_{n}e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{(4\pi t)^{n}}\left[
1+\sum\limits_{d=1}^{n-1}\frac{\gamma _{n-d-1}^{(\nu ,n)}(n-d-1)!}{(n-1)!}
t^{d}+\sum\limits_{d=n}^{+\infty }\left( \sum\limits_{p=0}^{n-1}\frac{\gamma
_{p}^{(\nu ,n)}}{(n-1)!}\Omega _{p}^{\left( \nu \right) }\left( d-n\right)
\right) t^{d}\right]
=\frac{\omega _{n}}{(4\pi t)^{n}}\left[ \sum\limits_{d=0}^{+\infty
}c_{d}^{\left( \nu ,n\right) }\ t^{d}\right] \left[ \sum\limits_{i=0}^{+
\infty }\frac{1}{i!}\left( \frac{n^{2}}{4}+\nu ^{2}\right) ^{i}\right]
where the coefficients c_{d}^{\left( \nu ,n\right) }are given by
c_{d}^{\left( \nu ,n\right) }=\left\{
\begin{array}{c}
\frac{(n-d-1)!}{(n-1)!}\gamma _{n-d-1}^{(\nu ,n)},\quad 0\leq d\leq n-1 \\
\sum\limits_{p=0}^{n-1}\frac{\gamma_{p}^{(\nu ,n)}}{(n-1)!}\Omega _{p}^{\left( \nu \right) }\left( d-n\right),\quad d\geq n
\end{array}
\right.
Since p\in \{0,\cdots, n-1\} in the second expression of (?), we may Write
\gamma _{p}^{(\nu ,n)}=\frac{(n-1)!}{p!}c_{n-p-1}^{\left( \nu ,n\right)}, \qquad 0\leq p \leq n-1
and then, the expression of c_{d}^{\left( \nu ,n\right) } can be simplify as follows
c_{d}^{\left( \nu ,n\right) }=\left\{
\begin{array}{c}
\frac{(n-d-1)!}{(n-1)!}\gamma _{n-d-1}^{(\nu ,n)},\quad 0\leq d\leq n-1 \\
\sum\limits_{p=0}^{n-1}\frac{c_{n-p-1}^{\left( \nu ,n\right)}}{p!}\Omega _{p}^{\left( \nu \right) }\left( d-n\right)
,\quad d\geq n.
\end{array}
\right.
This means that the coefficients c_{d}^{\left( \nu ,n\right) } satisfy a n-terms recurrence relation when d\geq n.
\left( i\right) Case even n\geq 2. In this case, the product (Proof) can also be presented as
\mathcal{\wp }'=\prod\limits_{l=\nu }^{\frac{n}{2}+\nu -1}\left(
r^{2}-l^{2}\right) \prod\limits_{s=1-\nu }^{\frac{n}{2}-\nu
-1}\left( r^{2}-s^{2}\right) \qquad \text{(product omitted for $n = 2$)}
and further it can be decomposed as
\mathcal{\wp }'=\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}r^{2p}.
where the numbers \tau _{0}^{(\nu ,n)},...,\tau _{n-1}^{(\nu ,n)} can be computed explicitly (see Sect.?). By using (?), the multiplicity in (Proof) takes the form
\dim _{\mathbb{C}}\mathcal{A}_{m}^{\nu }=\frac{2}{n!(n-1)!}
\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\left( m+\nu +\frac{n}{2}\right) ^{2p+1}
We now may insert the r.h.s of (?) into the trace formula (?) to get
\begin{split}
Tr\left( e^{\frac{1}{4}t\Delta _{\nu }}\right) &= e^{\left( \frac{n^{2}
}{4}+\nu ^{2}\right) t}\sum\limits_{m=0}^{+\infty }\left[ \frac{2}{n!(n-1)!}
\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\left( m+\nu +\frac{n}{2}\right)^{2p+1}\right] e^{-\left( m+\nu +\frac{n}{2}\right)^{2}t}\\
&=\frac{2 e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{n!(n-1)!}\sum\limits_{\mu=\nu+\frac{n}{2}}^{+\infty }e^{-\mu^{2}t}\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\mu^{2p+1} \\
&=\frac{ e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{n!(n-1)!}\left[ 2\sum\limits_{\mu=1}^{+\infty }e^{-\mu^{2}t}\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\mu^{2p+1}-2\sum\limits_{\mu=1}^{\nu+\frac{n}{2}-1}e^{-\mu^{2}t}\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\mu^{2p+1}\right] \\
&=\frac{ e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{n!(n-1)!}\left[S_1-S_2\right].
\end{split}
Now, recall the Jacobi’s theta function \left( \left[ ..\right] \text{, p..}
\right) :
\vartheta _{3}\left( t\right) :=2\sum\limits_{l=1}^{+\infty }le^{-l^{2}t}
Then one can see that the p derivative of \vartheta _{3}\left(
t\right) satisfies
\left( -1\right) ^{p}\left( \frac{d}{dt}\right) ^{p}\left[ \vartheta
_{3}(t)\right] =2\sum\limits_{l=1}^{+\infty }l^{2p+1}e^{-l^{2}t}
Therefore,
S_1=\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\left( -1\right)
^{p}\left( \frac{d}{dt}\right) ^{p}\left[ \vartheta _{3}(t)\right].
From (?) and (?) we see that for \mu raning over the set \{\nu ,\nu +1,\cdots ,\nu +\frac{n}{2}-1\} the sum
\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\mu ^{2p}=0,
and therefore S_2 reduces to
S_2=2\sum\limits_{\mu=1}^{\nu-1}e^{-\mu^{2}t}\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\mu^{2p+1}.
The above sum can be expressed as follows
\begin{split}
S_2&=2\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\sum\limits_{\mu=1}^{\nu-1}e^{-\mu^{2}t}\mu^{2p+1}\\
&=2\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}\sum\limits_{\ell =0}^{+\infty }\varrho _{p}^{(\nu )}\left( \ell \right) \
t^{\ell }
\end{split}
with coefficients
\varrho_{p}^{\left( \nu \right) }\left( \ell \right) =\frac{(-1)\ell }{\ell !
}\sum\limits_{\mu=1}^{\nu -1}\mu ^{2(p+\ell )+1}.
Now, we use the formula [prud1, p.596]:
\sum\limits_{k=1}^{m}k^q=\frac{1}{q+1}\left[ B_{q+1}(m+1)-B_{q+1}\right], \quad q=1,2,\cdots ,
for m=\nu-1, q=2(\ell+p)+1, to get
\varrho_{p}^{\left( \nu \right) }\left( \ell \right) =\frac{(-1)^{\ell}}{2(p+\ell +1)\ell !}\left[ B_{2(p+\ell +1)}\left(\nu\right)-B_{2(p+\ell +1)}\right].
Therefore, the trace formula becomes
\mathcal{T}:=Tr\left( e^{\frac{1}{4}t\Delta _{\nu }}\right) =\frac{e^{\left( \frac{n^{2}}{
4}+\nu ^{2}\right) t}}{n!(n-1)!}\sum\limits_{p=0}^{n-1}\tau _{p}^{(\nu ,n)}
\left[ \left( -1\right) ^{p}\left( \frac{d}{dt}\right) ^{p}\left[
\vartheta _{3}(t)\right] -\sum\limits_{\ell =0}^{+\infty }\frac{(-1)^{\ell}\left[ B_{2(p+\ell +1)}\left(\nu\right)-B_{2(p+\ell +1)}\right]}{(p+\ell +1)\ell !} \
t^{\ell } \right]
Applying now the asymptotic of the higher order derivative of \vartheta _{3}(t) as t\searrow 0^{+} [Awon2019, p.11] :
\left( \frac{d}{dt}\right) ^{\ell }\vartheta _{3}(t)\simeq \frac{(-1)^{\ell
}\ell !}{t^{1+\ell }}+\sum\limits_{j=\ell }^{+\infty }\frac{\left( -1\right) ^{j}B_{2j+2}}{(j+1)(j-\ell)!}t^{j-\ell }
where \left( B_{d}\right) are Bernoulli numbers in (?), substituting (?) into (?), we get
\begin{aligned}
\begin{split}
\mathcal{T} &\simeq \frac{e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{n!(n-1)!}\sum\limits_{p=0}^{n-1}\tau
_{p}^{(\nu ,n)}\left[\frac{p !}{t^{1+p}}+\left( -1\right)^{p}\sum\limits_{j=p}^{+\infty }\frac{\left( -1\right) ^{j}B_{2j+2}}{(j+1)(j-p)!}t^{j-p } -\sum\limits_{\ell =0}^{+\infty }\frac{(-1)^{\ell}\left[ B_{2(p+\ell +1)}\left(\nu\right)-B_{2(p+\ell +1)}\right]}{(p+\ell +1)\ell !} \ t^{\ell } \right]\\
&= \frac{e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{n!(n-1)!}\sum\limits_{p=0}^{n-1}\tau
_{p}^{(\nu ,n)}\left[\frac{p !}{t^{1+p}}+\sum\limits_{\ell=0}^{+\infty }\frac{\left( -1\right) ^{\ell}B_{2(\ell +p+1)}}{(\ell+p+1)\ell !}t^{\ell } -\sum\limits_{\ell =0}^{+\infty }\frac{(-1)^{\ell}\left[ B_{2(p+\ell +1)}\left(\nu\right)-B_{2(p+\ell +1)}\right]}{(p+\ell +1)\ell !} \ t^{\ell } \right] \\
&= \frac{e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{n!(n-1)!}\sum\limits_{p=0}^{n-1}\tau
_{p}^{(\nu ,n)}\left[\frac{p !}{t^{1+p}}+\sum\limits_{\ell =0}^{+\infty }\frac{(-1)^{\ell}\left[ 2B_{2(p+\ell +1)}-B_{2(p+\ell +1)}\left(\nu\right)\right]}{(p+\ell +1)\ell !} \ t^{\ell } \right]
\end{split}
\end{aligned}
We denote Vol\left(\textbf{P}^n(\mathbb{C})\right)=\omega_n=(4\pi)^n/n! and we use the fact that \tau_{n-1}^{(\nu ,n)}=1, to write
\begin{aligned}
\begin{split}
\mathcal{T} &\simeq \frac{\omega_n e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{(4\pi t)^n}\sum\limits_{p=0}^{n-1}\frac{\tau_{p}^{(\nu ,n)}}{(n-1)!}\left[p! t^{n-p-1}+\sum\limits_{\ell=0}^{+\infty }\frac{(-1)^\ell}{(p+\ell+1)\ell !}\left[ 2B_{2(p+\ell+1)}-B_{2(p+\ell+1)}(\nu)\right] \ t^{n+\ell } \right] \\
&= \frac{\omega_n e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{(4\pi t)^n}\left[1+\sum\limits_{p=0}^{n-2}\frac{\tau_{p}^{(\nu ,n)}p !}{(n-1)!} t^{n-p-1}+\sum\limits_{p=0}^{n-1}\frac{\tau_{p}^{(\nu ,n)}}{(n-1)!}\sum\limits_{\ell=0}^{+\infty }\frac{(-1)^\ell}{(p+\ell+1)\ell !}\left[ 2B_{2(p+\ell+1)}-B_{2(p+\ell+1)}(\nu)\right] \ t^{n+\ell } \right] \\
&= \frac{\omega_n e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{(4\pi t)^n}\left[1+T_1+T_2 \right]
\end{split}
\end{aligned}
For T_1, we have
T_1=\sum\limits_{p=0}^{n-2}\frac{\tau_{p}^{(\nu ,n)}p !}{(n-1)!} t^{n-p-1}=\sum\limits_{d=1}^{n-1}\frac{\tau_{n-d-1}^{(\nu ,n)}\left(n-d-1\right) !}{(n-1)!} t^{d}.
For T_2, we have
\begin{aligned}
\begin{split}
S_2 &=\sum\limits_{p=0}^{n-1}\frac{\tau_{p}^{(\nu ,n)}}{(n-1)!}\sum\limits_{\ell=0}^{+\infty }\frac{(-1)^\ell}{(p+\ell+1)\ell !}\left[ 2B_{2(p+\ell+1)}-B_{2(p+\ell+1)}(\nu)\right] \ t^{n+\ell }\\
&=\sum\limits_{p=0}^{n-1}\frac{\tau_{p}^{(\nu ,n)}}{(n-1)!}\sum\limits_{d=n}^{+\infty }\frac{(-1)^{d-n}}{(p+d-n+1)(d-n)!}\left[ 2B_{2(p+d-n+1)}-B_{2(p+d-n+1)}(\nu)\right] \ t^{d }\\
&=\sum\limits_{d=n}^{+\infty }\left(\sum\limits_{p=0}^{n-1}\frac{(-1)^{d-n}\tau_{p}^{(\nu ,n)}\left[ 2B_{2(p+d-n+1)}-B_{2(p+d-n+1)}(\nu)\right]}{(n-1)!(p+d-n+1)(d-n)!} \right) \ t^{d }
\end{split}
\end{aligned}
Therefore
\begin{aligned}
\begin{split}
\mathcal{T} &\simeq \frac{\omega_n e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right) t}}{(4\pi t)^n}\left[1+\sum\limits_{d=1}^{n-1}\frac{\tau_{n-d-1}^{(\nu ,n)}\left(n-d-1\right) !}{(n-1)!} t^{d}+\sum\limits_{d=n}^{+\infty }\left(\sum\limits_{p=0}^{n-1}\frac{(-1)^{d-n}\tau_{p}^{(\nu ,n)}\left[ 2B_{2(p+d-n+1)}-B_{2(p+d-n+1)}(\nu)\right]}{(n-1)!(p+d-n+1)(d-n)!} \right) \ t^{d } \right]
\end{split}
\end{aligned}
which can also be written as
\frac{\omega _{n}}{(4\pi t)^{n}}e^{\left( \frac{n^{2}}{4}+\nu ^{2}\right)
t}\sum\limits_{i=0}^{+\infty }c_{i}^{\left( \nu ,n\right) }t^{i}
where
\begin{aligned}
\begin{split}
c_{i}^{\left( \nu ,n\right) }=\left\{
\begin{array}{c}
1,\qquad\qquad\qquad\qquad\qquad\qquad\qquad i=0\\
\tau _{n-1-i}^{(\nu ,n)}\frac{\left(n-i-1\right) !}{\left( n-1\right) !},\qquad\qquad\qquad\qquad\qquad 1\leq i\leq n-1\\
\sum\limits_{p=0}^{n-1}\frac{(-1)^{i-n}\tau_{p}^{(\nu ,n)}\left[ 2B_{2(p+i-n+1)}-B_{2(p+i-n+1)}(\nu)\right]}{(n-1)!(p+i-n+1)(i-n)!} ,\qquad i\geq n
\end{array}
\right.
\end{split}
\end{aligned}
Since in the last expression of c_{i}^{\left( \nu ,n\right) } where i\geq n, we have p\in \{0,\cdots, n-1\} we may write
\gamma _{p}^{(\nu ,n)}=\frac{(n-1)!}{p!}c_{n-p-1}^{\left( \nu ,n\right)}, \qquad 0\leq p \leq n-1
and then, the expression of c_{i}^{\left( \nu ,n\right) } simplifies as follows
\begin{aligned}
\begin{split}
c_{i}^{\left( \nu ,n\right) }=\left\{
\begin{array}{c}
1,\qquad\qquad\qquad\qquad\qquad\qquad\qquad i=0\\
\tau _{n-1-i}^{(\nu ,n)}\frac{\left(n-i-1\right) !}{\left( n-1\right) !},\qquad\qquad\qquad\qquad\qquad 1\leq i\leq n-1\\
\sum\limits_{p=0}^{n-1}\frac{(-1)^{i-n}c_{n-p-1}^{\left( \nu ,n\right)}\left[ 2B_{2(p+i-n+1)}-B_{2(p+i-n+1)}(\nu)\right]}{p!(p+i-n+1)(i-n)!} ,\qquad i\geq n.
\end{array}
\right.
\end{split}
\end{aligned}
Therefore, we may rewrite (?) as
Tr\left( e^{\frac{1}{4}t\Delta _{\nu }}\right) \simeq \frac{\omega _{n}}{
(4\pi t)^{n}}\left[ \sum\limits_{k=0}^{+\infty }c_{k}^{\left( \nu ,n\right)
}\ t^{k}\right] \left[ \sum\limits_{j=0}^{+\infty }\frac{1}{j!}\left( \frac{
n^{2}}{4}+\nu ^{2}\right) ^{j}\right] .
Finally, in both cases the r.h.s of (?) and (?) may also be seen as a Cauchy product of two power series as
Tr\left( e^{\frac{1}{4}t\Delta _{\nu }}\right) \simeq\frac{1}{(4\pi t)^{n}}
\sum\limits_{d=0}^{+\infty }\left[ \omega _{n}\sum\limits_{i=0}^{d}\frac{
\left( n^{2}+4\nu ^{2}\right) ^{d-i}}{4^{d-i}\left( d-i\right) !}
c_{i}^{\left( \nu ,n\right) }\right] t^{d}.
This ends the proof of the theorem.